Академический Документы
Профессиональный Документы
Культура Документы
Submitted by:
Rice University and University of Houston
June, 2008
University of Houston
Department of Chemical Engineering
4800 Calhoun Street
Houston, TX 77204-4004
Prepared for:
United States Department of Energy
National Energy Technology Laboratory
Disclaimer .......................................................................................................... 4
Executive Summary ........................................................................................... 5
Background........................................................................................................ 8
Task 5: Carbon Inputs and Outputs to Gas Hydrate Systems ........................... 9
Approach................................................................................................. 9
Subtask 5.1: Constrain organic carbon inputs using iodine ......... 9
Subtask 5.2: Constrain methane outputs using authigenic minerals
(and carbon isotopes)................................................................. 11
Results and Discussion ........................................................................ 12
Subtask 5.1: Constrain organic carbon inputs using iodine ....... 12
Subtask 5.2: Constrain methane outputs using authigenic minerals
(and carbon isotopes)................................................................. 15
Conclusions........................................................................................... 18
References............................................................................................ 19
Task 6: Numerical Models for Quantification of Hydrate and Free Gas
Accumulations.................................................................................................. 21
Subtask 6.1 Model Development .......................................................... 21
Subtask 6.2: Conditions for existence of gas hydrate ........................... 24
Subtask 6.3 Compositional effects on BSR........................................... 25
Subtask 6.5: Processes leading to overpressure .................................. 30
Subtask 6.8: Sulfate profile as indicator of methane flux....................... 34
Students ................................................................................................ 36
Task 7: Analysis of Production Strategy .......................................................... 37
Code Comparison ..................................................................................... 37
Modeling of Warm Water Injection ............................................................ 41
Pore-Scale Model...................................................................................... 45
Conclusions............................................................................................... 46
Students .................................................................................................... 46
Task 8: Seafloor and Borehole Stability ........................................................... 47
Subtask 8.1: Sediment-Hydrate Properties ........................................... 47
Subtask 8.2: Modeling (In)stability ........................................................ 50
Subtask 8.3: Integrating geomechanical studies ................................... 51
Students ................................................................................................ 52
Hydrate Activities .................................................................................. 53
Acronyms and Abbreviations................................................................. 53
References............................................................................................ 54
Task 9: Geophysical Imaging of Gas Hydrate and Free Gas Accumulations... 55
Subtask 9.1: Preliminary processing and inversion of seismic data ...... 55
Task 10 Technology Transfer .......................................................................... 56
2
Cost Plan / Status ............................................................................................ 59
Milestone Plan / Status .................................................................................... 60
Appendix A: PhD Thesis of Gaurav Bhatnagar, “Accumulation of gas hydrates in
marine sediments” ........................................................................................... 65
Appendix B: “Effect of Overpressure on Gas Hydrate Distribution,” poster
presentation at the 6th International Conference on Gas Hydrate, Vancouver, BC,
July 6-10, 2008 .............................................................................................. 359
Appendix C: “Relating Gas Hydrate Saturation to Depth of Sulfate-Methane
Transition,” oral presentation at the 6th International Conference on Gas Hydrate,
Vancouver, BC, July 6-10, 2008. ................................................................... 368
3
Disclaimer
This report was prepared as an account of work sponsored by an agency
of the United States Government. Neither the United States Government nor any
agency thereof, nor any of their employees, makes any warranty, express or
implied, or assumes any legal liability or responsibility for the accuracy,
completeness, or usefulness of any information, apparatus, product, or process
disclosed, or represents that its use would not infringe privately owned rights.
Reference herein to any specific commercial product, process, or service by
trade name, trademark, manufacturer, or otherwise does not necessarily
constitute or imply its endorsement, recommendation, or favoring by the United
States Government or any agency thereof. The views and opinions of authors
expressed herein do not necessarily state or reflect those of the United States
Government or any agency thereof.
4
Executive Summary
The results of Phase 2 are presented as Tasks 5 through 9. With the
development of the numerical models, there have been increased interactions
between group members. Several articles and presentations have multiple
authors from both Chemical Engineering and Earth Science. Also Gangsheng
Gu of Chemical Engineering has been collaborating with Jerry Dickens of Earth
Science and Dr. Frederick S Colwell of Organ State University (Battelle Grant
BEA 00054031) on the effect of seafloor temperature on hydrate accumulation.
Task 5: Carbon Inputs and Outputs to Gas Hydrate Systems
We have determined the amount of iodine in sediment and pore waters
down boreholes at 10 locations, including 3 with gas hydrate. This work, when
combined with disparate data sets scattered throughout the literature, clearly
shows that iodine accumulates in marine sediment as a function of organic
carbon input over time. Iodine profiles provide a crucial constraint on the organic
carbon input to marine gas hydrate systems. We anticipate including iodine into
our modeling as a means to evaluate certain parameters. We have examined
carbon, sulfur and metal chemistry across the SMT at sites in the Japan Sea and
the Peru Trench to assess whether sulfate profiles can, in fact, be used to
determine the upward flux of methane. This appears to be a valid assumption in
the Sea of Japan, but only after all carbon fluxes are accounted for. In particular,
an upward flux of bicarbonate and carbonate precipitation impact geochemical
interpretations across the SMT. We are assessing whether this holds true for the
Peru Trench, and also, how this affects carbon isotopes. This information
provides crucial constraints for modeling the out of methane from marine gas
hydrate systems.
Task 6: Numerical Models for Quantification of Hydrate and Free Gas
Accumulations
The development of the numerical models for hydrate and free gas
accumulation has progressed to increase the fundamental understand of the
accumulation phenomena.
Subtask 6.1: Model development. A numerical model has been developed for
the simulation of the accumulation of hydrate and free gas over geological time
and length scales in one or two dimensions. Work has continued towards
extending the one-dimensional numerical model to two spatial dimensions.
Subtask 6.2: Conditions for existence of gas hydrate. The simulations
delineate basic modes of gas hydrate distribution in marine sediment, including
systems with no gas hydrate, gas hydrate without underlying free gas, and gas
hydrate with underlying free gas below the gas hydrate stability zone, for various
methane sources. The results are scaled using combinations of dimensionless
variables, particularly the Peclet number and Damkohler number, such that the
dependence of average hydrate saturation on numerous parameters can be
summarized using two contour maps, one for a biogenic source and one for
upward flux from a deeper source.
5
Our model presents a unified picture of hydrate accumulations that can be
used to understand well-characterized gas hydrate systems or to predict steady-
state average hydrate saturation and distribution at locations for which seismic or
core data are not available.
Subtask 6.3. Compositional Effect on BSR The presence of methane hydrate
is usually detected by a bottom simulating reflector (BSR). The BSR results from
the sudden change in acoustic impedance as the formation changes from being
hydrate saturated to free-gas saturated at the base of the hydrate stability zone.
If other hydrocarbons in addition to methane are present, this transition may take
place over a depth greater than the acoustic wavelength and the BSR may be
attenuated or absent.
The result of an example saturation calculation of the CH4 - C3H8 - H2O
hydrate system in the sediment is presented. It successfully demonstrated that
continuous change of SH and Sv over a long spatial distance (~300 m) is
possible. A gradual change of saturations, may result in gradual change of
acoustic properties, and induce weak BSR or even no BSR.
Subtask 6.5: Processes leading to overpressure. Work has continued through
one-dimensional numerical modeling to ascertain the factors and dimensionless
groups responsible for overpressure generation in gas hydrate systems. We had
determined through numerical simulations that the ratio of sediment absolute
permeability to the sedimentation rate was the key dimensionless group
controlling overpressure generation. The effect of overpressure, in turn, limits
the amount (thickness) of free gas that can accumulate below the GHSZ.
Hydrostatic pore pressures allow a relatively long connected gas column to form.
On the other hand, overpressure generation can significantly reduce the
thickness of this connected gas column before gas pressure reaches the
lithostatic limit at the BHSZ and causes sediments to fracture.
Subtask 6.8: Sulfate profile as indicator of methane flux. Numerical and
analytical models have been developed for inferring gas hydrate saturation in
marine sediments from pore water sulfate profiles. These models utilize the
depth of the sulfate-methane transition (SMT) as the primary input variable and
are valid for systems dominated by methane supply from deeper sources.
Results from these models are in agreement with gas hydrate saturations
estimated from resistivity logs/chloride data at several sites along Cascadia
Margin.
6
then only depressurization is the best method of production. At intermediate
injection temperature, the production rate changes non-monotonically with the
injection pressure. These parameters should be chosen carefully to optimize
recovery and recovery rate of gas. This paper addresses a very simple
homogeneous domain. Realistic reservoirs would have heterogeneity in
sediments as well as hydrate distribution, which need to be taken into account.
7
Background
A. Objective
This project seeks to understand regional differences in gas hydrate systems
from the perspective of as an energy resource, geohazard, and long-term climate
influence. Specifically, the effort will: (1) collect data and conceptual models that targets
causes of gas hydrate variance, (2) construct numerical models that explain and predict
regional-scale gas hydrate differences in 2- and 3-dimensions with minimal “free
parameters”, (3) simulate hydrocarbon production from various gas hydrate systems to
establish promising resource characteristics, (4) perturb different gas hydrate systems to
assess potential impacts of hot fluids on seafloor stability and well stability, and (5)
develop geophysical approaches that enable remote quantification of gas hydrate
heterogeneities so that they can be characterized with minimial costly drilling. Our
integrated program takes advantage of the fact that we have a close working team
comprised of experts in distinct disciplines.
The expected outcomes of this project are improved exploration and production
technology for production of natural gas from methane hydrates and improved safety
through understanding of seafloor and well bore stability in the presence of hydrates.
B. Scope of Work
The scope of this project is to more fully characterize, understand, and
appreciate fundamental differences in the amount and distribution of gas hydrate and
how this affects the production potential of a hydrate accumulation in the marine
environment. The effort will combine existing information from locations in the ocean
that are dominated by low permeability sediments with small amounts of high
permeability sediments, one permafrost location where extensive hydrates exist in
reservoir quality rocks and other locations deemed by mutual agreement of DOE and
Rice to be appropriate. The initial ocean locations are Blake Ridge, Hydrate Ridge, Peru
Margin and GOM. The permafrost location is Mallik. Although the ultimate goal of the
project is to understand processes that control production potential of hydrates in marine
settings, Mallik will be included because of the extensive data collected in a producible
hydrate accumulation. To date, such a location has not been studied in the oceanic
environment. The project will work closely with ongoing projects (e.g. GOM JIP and
offshore India) that are actively investigating potentially economic hydrate accumulations
in marine settings.
The overall approach is fivefold: (1) collect key data concerning hydrocarbon
fluxes which is currently missing at all locations to be included in the study, (2) use this
and existing data to build numerical models that can explain gas hydrate variance at all
four locations, (3) simulate how natural gas could be produced from each location with
different production strategies, (4) collect new sediment property data at these locations
that are required for constraining fluxes, production simulations and assessing sediment
stability, and (5) develop a method for remotely quantifying heterogeneities in gas
hydrate and free gas distributions. While we generally restrict our efforts to the locations
where key parameters can be measured or constrained, our ultimate aim is to make our
efforts universally applicable to any hydrate accumulation.
8
Task 5: Carbon Inputs and Outputs to Gas Hydrate Systems
Approach
The amount and distribution of gas hydrate in marine sediment depends on
several factors. Our project-related modeling efforts [Bhatnager et al., 2007a,
2008], as well as results from other studies [e.g., Davie & Buffett, 2001], indicate
two particularly important factors: (1) the sedimentary input of labile organic
carbon (Corg) over time, and (2) the output of gas via anaerobic oxidation of
methane (AOM). We are constraining these factors by generating key chemical
data sets using sediment and pore water obtained from present-day gas hydrate
systems.
9
Figure 5.1: Sediment and pore water iodine concentrations in drill holes from gas
hydrate systems on the Peru Margin (685, 688, 1230) and Blake Ridge (997).
10
Subtask 5.2: Constrain methane outputs using authigenic minerals
A major loss of methane from all gas hydrate systems is AOM in shallow
sediment (Fig. 5.2). In fact, this may be the dominant methane output from most
systems, and its flux may directly link to the abundance of underlying gas hydrate
[Bhatnagar et al., 2008]. In general, and at a sulfate-methane transition (SMT),
upward moving methane reacts with downward diffusing sulfate, presumably in a
1:1 relationship: CH4 + SO42- ---> H2O + HCO3- + HS-. Because in situ methane
concentrations and fluxes can be difficult to measure, several workers have
suggested that upward methane fluxes could be determined from dissolved
sulfate (or bicarbonate) fluxes.
Several publications have provided data in support of this idea [Borowski et
al., 1999; Niewöhner et al., 1998; Luff and Wallmann, 2003]. However, other
publications have argued that sulfate and methane fluxes do not balance across
the SMT [Fossing et al., 2000; Aharon and Fu, 2003; Joye et al., 2004; Berelson
et al., 2005]. In particular, downward sulfate fluxes appear much greater than
upward methane fluxes. This has led to current debate as to whether pore water
sulfate profiles actually can be used to constrain methane fluxes. Moreover, it is
not obvious that present-day sulfate profiles above gas hydrate systems are
legitimate for constraining methane fluxes over the time-scales of interest to
modeling [e.g., Dickens, 2001]. These issues need clarification, as the sulfate
profiles provide a crucial model parameter for understanding gas hydrate
abundance [Bhatnager et al., 2008].
11
Figure 5.2: Schematic of dissolved solute behavior in shallow sediment above
gas hydrate systems [from Snyder et al., 2007]. Also shown are the sulfate-
methane transition (SMT) and the location of authigenic mineral formation.
AOM = anaerobic oxidation of methane.
We proposed to generate metal profiles in pore water and sediment across
the SMT at several locations with underlying gas hydrate. Our goal is to evaluate
the relationship between sulfate and methane profiles at present and in the past,
and then use established methane fluxes to constrain our numerical models for
gas hydrate formation. Previous work [e.g., Dickens, 2001; Aloisi et al., 2003;
Luff and Wallmann, 2003] has indicated that authigenic minerals (carbonate and
barite), and their constiituent elements (Ca, Mg, Sr, Ba), can be used to constrain
carbon and sulfur fluxes across the SMT.
12
Iodine contents of sediment samples were determined using an established
technique. First, iodine was released from weighed aliquots of sediment to a trap
solution by hydropyrolysis. Approximately 200 mg of powdered sample was
placed in a porcelain boat, combined with vanadium pentoxide catalyst, and
heated to 1100°C in a tube furnace using a quartz process tube while being
exposed to a steady stream of wet oxygen gas. Vapors formed were passed
through a reducing solution of 0.1% tetramethyammonium hydroxide trap
solution. Solutions were analyzed for I concentration by inductively couple
plasma mass spectrometry (ICP-MS), using mass 127 and rhodium (mass 103)
as an internal standard. Pore fluids were diluted 1:500 wt % with 18MΩ water, to
which tetramethylammonium hydroxide and the Rh internal standard were
added. These were analyzed for I concentrations by ICP-MS as noted above.
When combined with previously published data, our results allow us to
make generalizations regarding the distribution and accumulation of marine
iodine across a variety of depositional environments. The extreme pore water
iodide concentrations found on Blake Ridge are not unique (Fig. 5.1), but
characterize large gas hydrate systems elsewhere (e.g., Hydrate Ridge, Peru
Margin). On the other hand, sites with minimal organic carbon input over time
have no appreciable iodine (Fig. 5.3). This is true for sites that do not receive
high amounts of organic carbon, are too young, or both. Our work further shows
that sedimentary iodine contents are only high near the seafloor. A somewhat
amazing summary is that most of the iodine on Earth (>50%) occurs in methane-
charged sediment along continental slopes, and much of this resides in pore
fluids at sub-seafloor depths greater than 25 meters rather than in sediment.
Clearly, iodine is accumulating over time and represents an integrated signal of
organic input, likely because of the cycling hypothesized above.
We are writing these results and expect to have a paper submitted by the
end of summer. We will then attempt to incorporate iodine into our numerical
modeling. We note at this juncture, though, that some specific sites with gas
hydrate, while conforming to our general views, are problematic because of
external (deep) iodine sources. It may not be easy to use iodine at sites with high
rates of fluid advection (e.g., Gulf of Mexico).
13
Figure 5.3: Pore water iodine concentrations in drill holes from the Peru Shelf
and Equatorial Pacific where no gas hydrate occurs. Note scale change for the
x-axis compared to Figure 5.1.
14
Subtask 5.2: Constrain methane outputs using authigenic minerals (and carbon
isotopes)
We were tentatively awarded the grant at the end of 2006 but did not
receive full confirmation or funding until late summer of 2007. To initiate Task 5
and to better understand methane outputs, therefore, we generated a suite of
basic chemical data across the SMT using samples that we had already collected
as part of a project with Japanese colleagues. These samples come from
Umitaka Spur (Japan Sea), a location where shallow sediment contains
abundant gas and gas hydrate. The data produced were Ca, Mg, Sr, S, and Ba
concentrations in both pore water and sediment. These were determined using
an Inductively Coupled Plasma Atomic Emission Spectrometer (ICP-AES)
following procedures and extractions detailed by Snyder et al. [2007].
When combined with pore water HCO3- data generated on ship, problems
and answers to carbon (and sulfur) cycling across the SMT become clear. A
major issue is that very few studies have tried to constrain all relevant fluxes. In
particular, across the SMT, AOM is not the only sink and source of carbon; there
are three additional fluxes of interest: dissolved HCO3- rising into the SMT from
below, excess dissolved HCO3- leaving the SMT to the seafloor, and precipitation
of authigenic carbonate in sediment at (or near) the SMT. All five fluxes can be
calculated for the Japan Sea cores (and in many cores) using our pore water
data and standard diffusion equations (Fig. 5.4). These calculations indicate that
sulfate and methane fluxes do balance across the SMT. However, some of the
HCO3- entering and leaving the SMT comes from deep in the sediment column
not from AOM (Fig. 5.4). Authigenic carbonate fronts were also identified at or
near the SMT in all three cores. Given the dissolved fluxes of Ca, Mg, Sr and
HCO3- into these fronts and the amounts of carbonate hosted by them further
suggests that the SMT has been close to that at present-day for a long time
(>20,000 years). These results have now been published [Snyder et al., 2007].
15
Figure 5.4: Pore water (points) and solid phase (shades) elemental profiles
across the SMT in shallow sediment from three cores in the Japan Sea
[Snyder et al., 2007]. Also shown are fluxes in mol/m2-kyr.
16
We have now collected and analyzed the sediment metals contents across
a site with gas hydrate located in the Peru Trench (ODP Site 1230). This site was
chosen because we have already published very detailed pore water records
[Donohue et al., 2006], because it is a very different environment from the Japan
Sea, and because, being a deep borehole, it is amenable to our numerical
modeling [Bhatnager et al., 2007, 2008, in prep]. We have determined that there
is a 2-m thick horizon with high amounts of authigenic carbonate (calcite) and
barite across the SMT (Fig. 5.5). Similar to cores from the Japan Sea, this
horizon attests to a methane output that has been similar to present-day over a
long time interval (i.e., the system is close to steady-state, at least over the
>10,000 yrs). We are presently modeling the fluxes of constituents into the SMT,
to establish whether methane and sulfate fluxes balance here also, once other
carbon and sulfur fluxes are considered. We plan on submitting this work for
publication by the end of summer.
One argument that has been given for carbon balance inequity across the
SMT concerns stable carbon isotopes of dissolved HCO3- and authigenic
carbonate. It has been suggested that d13C values much “heavier” than methane
(e.g., -25‰ instead of -60‰) at the SMT indicate consumption of sulfate by other
processes. To test this, we have also collected and analyzed samples for carbon
isotopes. These measurements give values of -12‰ to -5‰, much “heavier” than
if all HCO3- was derived from AOM. However, we suspect that this is because
there is a very large upward flux of HCO3- from depth, so that methane and
sulfate fluxes may in fact still balance. We also plan on submitting this work for
publication by the end of summer, either with the above manuscript or
separately. Lastly, we have modeled the abundance of gas hydrate at this
location [Bhatnager et al., in preparation]. The results of our sediment chemistry
work will enable us to evaluate whether key model parameters are reasonable.
17
Figure 5.5. Pore water and solid phase chemistry across the SMT at ODP Site
1230, a borehole in the Peru Trench with gas hydrate.
Conclusions
Models concerning the abundance and distribution of gas hydrate in marine
sediment require constraints on carbon inputs and outputs, fluid flow and
temporal evolution. Iodine appears to be a proxy for the integrated flux of organic
carbon over time, such that sites that have accumulated large amounts of
organic carbon have extreme iodine abundance. Carbon, sulfur and metal
profiles can be used to assess whether sulfate profiles are good proxies for
upward methane fluxes, and whether these fluxes have changed over time. Our
chemical analyses of sediment appear to be providing us reasonable constraints
that we can use in our models.
18
References
Aloisi, G., K. Wallmann, S.M. Bollwerk, A. Derkachev, G. Bohrmann, & E. Suess,
2004. The effect of dissolved barium on biogeochemical processes at cold
seeps. Geochimica et Cosmochimica Acta, 68: 1735-1748.
Aharon, P. & B. Fu, 2003. Sulfur and oxygen isotopes of coeval sulfate-sulfide in
pore fluids of cold seep sediments with sharp redox gradients. Chemical
Geology, 195: 201-218.
Berelson, W.M., M. Prokopenko, F.J. Sansone, A.W. Graham, J. McManus &
J.M. Bernhard, 2005. Anaerobic diagenesis of silica and carbon
incontinental margin sediments: discrete zone of TCO2 production.
Geochimica et Cosmochimica Acta, 69: 4611-4629.
Borowski, W.S., C.K. Paull & W. Ussler III, 1999. Global and local variations of
interstitial sulfate gradients in deep-water, continental margin sediments:
Sensitivity to underlying methane and gas hydrates. Marine Geology, 159:
131-154.
Davie, M.K. & B.A. Buffett, 2001. A numerical model for the formation of gas
hydrate below the seafloor. J. Geophysical Research, 106: 497-514.
Dickens, G.R., 2001. Sulfate profiles and barium fronts in sediment on the Blake
Ridge: Present and past methane fluxes through a large gas hydrate
reservoir. Geochimica et Cosmochimica Acta, 65: 529-543
Donohue, C.M., G.R. Dickens & G.T. Snyder, 2006. Data Report: Major cation
concentrations of interstitial waters collected from deep subsurface
microbial communities (ODP Leg 201). Proceedings ODP, Scientific
Reports, 201: available http://www-
odp.tamu.edu/publications/201_SR/104/104.htm>
Egeberg, P.K. & G.R. Dickens, 1999. Thermodynamic and pore water halogen
constraints on gas hydrate distribution at ODP site 997 (Blake Ridge).
Chemical Geology, 153: 53-79.
Fossing, H., T.G. Ferdelman & P. Berg, 2000. Sulfate reduction and methane
oxidation in continental margin sediments influenced by irrigation (South-
East Atlantic off Namibia). Geochimica et Cosmochimica Acta, 64: 897-910.
Joye, S.B., A. Boetius, B.N. Orcutt, J.P. Montaya, H.N. Schulz, M.J. Erickson &
S.K. Lugo, 2004. The anaerobic oxidation of methane and sulfate reduction
in
sediments from Gulf of Mexico cold seeps. Chemical Geology, 205: 219-
238.
Luff, R. & K. Wallmann, 2003. Fluid flow, methane fluxes, carbonate precipitation
and biogeochemical turnover in gas hydrate-bearing sediments at Hydrate
Ridge, Cascadia Margin: Numerical modeling and mass balances.
Geochimica et Cosmochimica Acta, 67: 3403-3421.
19
Niewöhner, C., C. Hensen, S. Kasten, M. Zabel & Schulz, H.D., 1998. Deep
sulfate reduction completely mediated by anaerobic methane oxidation in
sediments of the upwelling area off Namibia. Geochimica et Cosmochimica
Acta, 62: 455-464.
Snyder, G.T., A. Hiruta, R. Matsumoto, G.R. Dickens, H. Tomaru, R. Takeuchi, J.
Komatsubara, Y. Ishida & H. Yu, 2007. Pore water profiles and authigenic
mineralization in shallow marine sediments above the methane-charged
system on Umitaka Spur, Japan Sea. Deep-Sea Research (II), 54: 1216-
1239.
20
Task 6: Numerical Models for Quantification of Hydrate and Free Gas
Accumulations
Subtask 6.1: Model development. A numerical model has been developed for
the simulation of the accumulation of hydrate and free gas over geological time
and length scales in one or two dimensions. The details of the model are
described by the PhD thesis of Gaurav Bhatnagar, which is included in Appendix
A of this report. The mechanism included in this model include methanogenesis,
sedimentation and compaction, advection and diffusion, sulfate-methane
transition, Darcy’s law, and constitutive relations for porosity, permeability,
relative permeability, and capillary pressure, and a thermodynamic model for the
gas/liquid/hydrate equilibrium of methane and brine. The results of Task 6 are
based on this model.
Work has continued towards extending the one-dimensional
numerical model to two spatial dimensions. Upward free gas migration due to
buoyancy has also been included in the model. We have also developed the
capability to model the effect of heterogeneities in focusing fluid flow and
concentrating gas hydrate/free gas in two dimensions. We present two simple
test cases to illustrate how gas hydrate/free gas is concentrated along high
permeability conduits.
The first case models a system with a single vertical fracture located along
the center of the grid that extends from the seafloor down to the bottom of the
simulation domain. The fracture permeability is assumed to be 100 times greater
than that of the surrounding sediment. Over geologic time, this fracture gets
buried away from the seafloor with the downward moving sediment. Figure 6.1-1
shows the position of the fracture at a later time and the effect of focused fluid
flow on gas hydrate and free gas saturation contours. It can be clearly seen from
Figure 6.1-1 that gas hydrate as well as free gas is concentrated within and
around the fracture compared to the surrounding sediment.
Figure 6.1-1: Gas hydrate (left) and free gas (right) saturation contours at
dimensionless time t = 0.1 after the fracture is introduced in the system. Dashed
lines show the position of the fracture within the system.
21
Figure 6.1-2 shows the evolution of gas hydrate and free gas saturation at
a later time ( t = 0.5). Peak gas hydrate saturation occurs within the fracture and
close to the base of the GHSZ (~20%), which is almost twice the value in
surrounding sediments at the same normalized depth. Free gas also
accumulates in greater amount within the fracture, with peak saturation of about
50% beneath the GHSZ. Compared to Figure 6.1-1, the fracture has moved
down to about half of the depth of the GHSZ. Consequently, the gas hydrate
saturation in the upper half of the GHSZ becomes relatively homogeneous. At
later times, the fracture gradually moves out of the GHSZ causing sediments to
have a much more homogeneous hydrate distribution within the entire GHSZ.
Figure 6.1-2: Gas hydrate (left) and free gas (right) saturation contours at
dimensionless time t = 0.5 after the fracture is introduced in the system. Dashed
lines show the position of the fracture within the system.
22
At a later time ( t =0.75), the sand layer almost passes completely through
the GHSZ (Figure 6.1-4). Consequently, hydrate saturation returns to a more
homogeneous distribution within the GHSZ. Free gas saturation increase to
about 60% within the sand layer just below the GHSZ and also migrates laterally
to increase peak gas saturation in the lower permeability sediments to about 30%
(Figure 6.1-4).
The above simulations were relatively simple test cases performed to
validate our two-dimensional model and code. Effect of different system
parameters, such as thickness of beds, permeability contrasts, dip angles, and
combination of different permeability beds with fracture networks are planned as
future work.
Figure 6.1-3: Gas hydrate and free gas saturation contours at dimensionless
time t =0.25 after deposition of the sand layer within low permeability clay
sediments. Peak hydrate saturation within the sand layer increases to about
30%, but the color axis is scaled to a maximum of 15% to show the other
contours more clearly. Vertical exaggeration is about 2:1.
23
Figure 6.1-4: Gas hydrate and free gas saturation contours at dimensionless
time t =0.75. Sand layer almost passes through the GHSZ, causing hydrate
saturation to become more homogeneous laterally within the GHSZ. Vertical
exaggeration is about 2:1.
In conclusion, numerical models have been developed in 1-D and 2-D for
the accumulation of hydrate and free-gas. The applications of this model are
described in the following subtasks.
Subtask 6.2: Conditions for existence of gas hydrate.
G. Bhatnagar, W. G. Chapman, G. R. Dickens, B. Dugan, and G. J. Hirasaki,
“Generalization Of Gas Hydrate Distribution and Saturation In Marine
Sediments by Scaling of Thermodynamic and Transport Processes,”
American Journal of Science, Vol. 307, June, 2007, P. 861–900, DOI
10.2475/06.2007.01
ABSTRACT.
Gas hydrates dominated by methane naturally occur in deep marine
sediment along continental margins. These compounds form in pore space
between the seafloor and a sub-bottom depth where appropriate stability
conditions prevail. However, the amount and distribution of gas hydrate within
this zone, and free gas below, can vary significantly at different locations. To
understand this variability, we develop a one-dimensional numerical model that
simulates the accumulation of gas hydrates in marine sediments due to upward
and downward fluxes of methane over time. The model contains rigorous
thermodynamic and component mass balance equations that are solved using
expressions for fluid flow in compacting sediments. The effect of salinity on gas
hydrate distribution is also included.
24
The simulations delineate basic modes of gas hydrate distribution in
marine sediment, including systems with no gas hydrate, gas hydrate without
underlying free gas, and gas hydrate with underlying free gas below the gas
hydrate stability zone, for various methane sources. The results are scaled using
combinations of dimensionless variables, particularly the Peclet number and
Damkohler number, such that the dependence of average hydrate saturation on
numerous parameters can be summarized using two contour maps, one for a
biogenic source and one for upward flux from a deeper source. Simulations also
predict that for systems at steady state, large differences in parameters like
seafloor depth, seafloor temperature and geothermal gradient cause only small
differences in average hydrate saturation when examined with scaled variables,
although important caveats exist. Our model presents a unified picture of
hydrate accumulations that can be used to understand well-characterized gas
hydrate systems or to predict steady-state average hydrate saturation and
distribution at locations for which seismic or core data are not available.
(Only the abstract is presented here since the manuscript has been
published.)
where ni is the amount of species i in the system (unit: mol), i = CH4, C3H8, H2O.
25
The water free molar fraction of species i is denoted as:
ni xi
xiwf = = , i =C H4, C3H8.
nCH 4 + nC 3 H 8 xCH 4 + xC 3 H 8
Fig. 6.3-1 shows the incipient hydrate formation pressure of the CH4-C3H8-
H2O System. For xCwf3 H 8 = 0.01, the incipient hydrate formation pressure differs
very much from that for the pure CH4 system (i.e., xiwf = 0).
26
Phase Regions for the CH4-C3H8-H2O System
0
1
Water-free
2 C3 mol Fraction
0.05
3
Pure CH4
4 Region C:
P, MPa
Aq+V
5
6 Region B:
Aq+sII+V
7 Region A:
Aq+H (=sI+sII) SH=0
8
9 Sv =0
10
276 278 280 282 284 286 288 290 292 294 296
T, K
Fig. 6.3-2 presents the phase regions of the CH4-C3H8-H2O System ( xCwf3 H 8
is 0.05). 3 phase regions are marked in the figure. In Region A, both sI and sII
hydrates are stable, while in Region B and C, sI is not stable. In Region B, sII is
stable, while in Region C, sII hydrate is not stable. Therefore, in Region B, 3
phases can co-exist: Aq, H, and V. The boundaries for Sv=0 and SH=0 are
marked in the Fig. 6.3-2.
27
(2) Gradual Phase Transition (i.e. Saturation Change) in Sediment
Phase Regions for the CH4-C3H8-H2O System
0
1
Water-free
2 C3 mol Fraction
0.05
3
Pure CH4
4
Seafloor
P, MPa
A
5
Z on
6 eA B
: Aq Zo n
+H eC
(sI+ : Aq
7 sII) Geotherm +V
8 Sv =0 Z on
eB
: Aq SH=0
+ sI
9 I+ V
C D
10
276 278 280 282 284 286 288 290 292 294 296
T, K
28
Saturation Profile
0 5
50 Zone A: Aq + H (sI+sII)
sI sII All Hydrate
S =0 (=sI+sII)
100 V 6
z, Depth below Seafloor (mbsf)
150
A B
200 7
V
P (MPa)
250
300 Tsf=276.15 K 8
Zone B: Aq+H+V
G=0.04 K/m
350
Overall Composition
400 x CH4=0.019 9
x C3H8=0.001
450
C D
Zone C: Aq+V S =0
H
500
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18
Saturation
29
A gradual saturation change will result in the gradual change of acoustic
properties with increase in depth, and consequently, very possibly induce a weak
BSR, or even absence of BSR.
Conclusion
(1) For the CH4-C3H8-H2O hydrate system, the incipient hydrate formation
conditions are presented. There is a big difference for the incipient hydrate
formation condition of the CH4-C3H8-H2O hydrate system, from that of the CH4-
H2O hydrate system, even when the water-free-propane molar fraction is only
0.01.
(2) 3 different phase regions are described for different P-T conditions. As
shown in Fig. 6.3-2. Region B is especially important, because Aq, H (sII), V can
co-exist. Therefore, in the sediment, 3 zones can be present. Zone B, is the
phase-transition-zone, because Aq, H (sII), V co-exist, and SH and Sv change
gradually.
Future Work
The work presented in this report is based on constant composition in
sediment as an example. In real situations, the composition in spatial domain is
not constant, but dependent on various kinds of factors, such as the fluid flow,
diffusion, and phase transformation. Therefore, compositional fluid migration
simulation with consideration of fluid flow, diffusion, phase transformation, will be
carried out to compute realistic compositions and saturation profiles. The
acoustic impedance and seismic reflection will be computed from the saturation
profile.
30
where k0 is the sediment permeability, ρ w is seawater density, μ w is viscosity
and S is sedimentation rate. Figure 6.5-1 shows that higher values of Nsc
indicate high permeability and/or low sedimentation rate, leading to hydrostatic
pore pressures. As Nsc decreases, pore pressure starts to increase towards
lithostatic values.
Figure 6.5-1: Effect of the dimensionless group Nsc on pore pressure evolution.
Plith, Pw and Phydro denote lithostatic, pore pressure and hydrostatic pressure
profiles, respectively. Relatively higher values of Nsc lead to almost hydrostatic
pore pressure, while Nsc close to unity results in pore pressure that is very close
to the lithostatic limit.
The effect of overpressure, in turn, limits the amount (thickness) of free
gas that can accumulate below the GHSZ. This situation is depicted
schematically in Figure 6.5-2a, where hydrostatic pore pressures allow a
relatively long connected gas column to form. On the other hand, Figure 6.5-2b
shows that overpressure generation can significantly reduce the thickness of this
connected gas column before gas pressure reaches the lithostatic limit at the
BHSZ and causes sediments to fracture.
31
Figure 6.5-2: Schematic illustration of the effect of overpressure on the
maximum thickness of the connected free gas column beneath the GHSZ. Plith,
Pw, Phydro and Pg denote lithostatic, pore pressure, hydrostatic and gas pressure
profiles, respectively. Development of overpressure can significantly reduce the
thickness of the connected gas column before fracturing occurs.
This work has been accepted for a poster presentation at the 6th
International Conference on Gas Hydrates, Vancouver, British Columbia, 2008.
The abstract for this presentation is titled “Effect of Overpressure on Gas
Hydrate Distribution”. (Appendix B)
32
Figure 6.5-3: Normalized pressure profiles (left) and gas hydrate and free gas
saturation profiles (right) for Nsc=1000. Gas pressure at this time is just equal to
the lithostatic stress. A relatively thick connected gas column exists beneath the
GHSZ at this state.
Figure 6.5-4: Normalized pressure profiles (left) and gas hydrate and free gas
saturation profiles (right) for Nsc=10. Gas pressure at this time is just equal to the
lithostatic stress. Compared to Figure 6.5-3, a relatively thin connected gas
column exists beneath the GHSZ at this state.
33
Subtask 6.8: Sulfate profile as indicator of methane flux.
Numerical and analytical models have been developed for inferring gas
hydrate saturation in marine sediments from pore water sulfate profiles. These
models utilize the depth of the sulfate-methane transition (SMT) as the primary
input variable and are valid for systems dominated by methane supply from
deeper sources. Results from these models are in agreement with gas hydrate
saturations estimated from resistivity logs/chloride data at several sites along
Cascadia Margin.
The numerical model is explained in a short article in Geophysical
Research Letters, titled “Sulfate-methane transition as a proxy for average
methane hydrate saturation in marine sediments”. Geophysical Research
Letters, VOL. 35, L03611, doi:10.1029/2007GL032500, 2008. The details are in
the thesis in Appendix A.
Analytical theory has also been developed to predict steady-state gas
hydrate saturation in deep-source systems using the depth of the SMT as the
primary input. This theory allows calculation of the gas hydrate saturation profile,
as well as the sulfate and methane concentration profiles, using simple analytical
expressions. Figure 6.8-1 below shows gas hydrate saturation profiles as a
function of scaled depth below the seafloor for different values of the SMT depth
Ls , which is the ratio of the SMT depth to the depth to the base of the gas
hydrate stability zone (GHSZ). Shallow SMT depths indicate higher methane flux
and, consequently, higher gas hydrate saturation. Results from our numerical
models (crosses) compare favorably with our analytical results (curves).
Several interesting aspects of gas hydrate systems can also be explained
using our analytical theory. For example, we show why the depth to the first
occurrence of gas hydrate below the seafloor is often 10-12 times the depth of
the SMT (Figure 6.8-2) using our analytical theory. This “10 x SMT” relationship
has been often hypothesized in the literature based on field observations. Figure
6.8-2 shows this ratio as a function of scaled SMT depth and demonstrates that
the ratio is close to 10-12 for relatively large SMT depths, but increases to higher
values for relatively shallower SMT depths.
Conclusions The depth of the sulfate-methane transition is an indicator of the
average hydrate saturation when the accumulation is the result of steady-state
accumulation from deeper sources.
This analytical theory has been written into a longer manuscript during this
quarter and will be submitted shortly for publication in Geochemistry,
Geophysics, Geosystems. This work has also been accepted for oral
presentation at the 6th International Conference on Gas Hydrates, Vancouver,
British Columbia, 2008. The abstract for this presentation is titled “Relating Gas
Hydrate Saturation to Depth of Sulfate-Methane Transition”. (Appendix C)
34
Figure 6.8-1: Steady-state gas hydrate saturation profiles for different scaled
SMT depths Ls . Crosses denote numerical model results, while curves
represent the analytical model.
Figure 6.8-2: Relationship between the ratio of depth to the first occurrence of
gas hydrate ( Lt − Lh ) to the SMT depth ( Ls ) as a function of the scaled SMT
depth ( Ls / Lt ) for two different seafloor depths.
35
Task 6 Students
Gaurav Bhatnagar completed his Chemical Engineering PhD requirements and
began employment in February 2008 with Shell Oil Company at the Westhollow
Research Center.
Guangsheng Gu is continuing and Sayantan Chatterjee has been added as
Chemical Engineering PhD students.
36
Task 7: Analysis of Production Strategy
We are moving forward on Task 7 as scheduled and as outlined by the
Milestone Status. Task 7.1 involves code comparison and pore-scale
development. Code comparison is complete and our results have been
communicated Prof. Brian Anderson of University of West Virginia. Pore-scale
modeling is expected to be completed in June. Petrophysical and
thermophysical properties will be computed in the next 12 months. Task 7.2
involves modeling production strategies and incorporating reservoir
heterogeneities. These subtasks would be executed in the following years. We
are working with Prof. Hirasaki’s group (Task 6) and Dr. Dugan’s group (Task 8)
to estimate the hetergeneities in reservoir properties. We are presenting a paper
in the upcoming ICGH in Vancouver and staying appraised of the industrial and
academic perspectives.
Overview
Task 7.1a includes code comparison and understanding of hydrate distribution at
the pore-scale. We have conducted the following in the last year. We
participated in the NETL methane hydrate code comparison study to evaluate the
capabilities of the in-house (University of Houston) simulator with respect to other
existing hydrate simulators participating in the code comparison study, we
simulated warm water injection into simple homogeneous hydrate reservoirs and
depressurization to identify the saturation history pores go through during hydrate
dissociation. We performed simulations of hydrate formation in simple pores
using the University of Houston in-house simulator.
Approach
We have completed simulation of the first four problems set up by the Code
Comparison Study group by our in-house hydrate simulator. Our results have
been communicated to Prof. Brian Anderson, the coordinator of the Code
Comparison Study group. Warm water injection into a homogeneous hydrate
reservoir has been modeled. Hydrate formation in simple cylindrical pores have
been modeled.
Problem 1
This problem calculates transition to equilibrium in the absence of hydrates. The
aqueous saturation, aqueous pressure and temperature calculated by our
simulator are in complete agreement with the other results. There is a slight
discrepancy in the mass fraction of methane in the aqueous phase. This may be
due to different correlations used in different simulators for methane solubility.
Our results are labeled UH in Figure 7-1.
37
UH 1
8.00E-01
TOUGH 10
UH 100
HRS 100
6.00E-01
MH21 100
STARS 100
STOMP 100
4.00E-01
TOUGH 100
UH 1000
HRS1000
2.00E-01
MH21 1000
STARS 1000
STOMP 1000
0.00E+00
TOUGH 1000
0 5 10 15 20 25
UH 10000
Grid Block Number HRS 10000
Problem 2
This problem calculates transition to equilibrium in a closed domain with hydrate
dissociation. There is a discrepancy in the position of the saturation front (Figure
7-2). We believe, that is due to the unspecified correlation of change of
permeability with hydrate saturation. (This issue is present in all the problems
except Problem 1). The sediment permeability changes with hydrate saturation;
we have used a correlation. We do not know the correlations used by others.
Heat conductivity of the hydrates is also not specified. It affects the heat flow
and hence the slight deviation in the results. Our results are shown in yellow in
the plots.
38
Hydrate Saturation at 100 Days
0.7
0.6 UH_100
0.5 HRS_100
0.4 MH_100
SH
0.3 STARS_1000
0.2 STOMP_100
0.1 T OUGH_100
0
0 5 10 15 20 25
Gr id Blo ck Numb er
Problem 3
This problem calculates hydrate dissociation in an open 1-D domain. In the first
part, hydrate is dissociated by thermal stimulation. In the second part, hydrate is
dissociated by depressurization without ice formation. In the third part, hydrate is
dissociated by depressurization with ice formation for a very low pressure.
0.9
0.8
0.7 UH_5days
0.6 Tough_5days
0.5 STOMP_5days
SH
0.4 MH21_5days
0.3 HRS_5days
0.2
0.1
0
0 5 10 15 20 25 30 35
39
Figure 7-3. Comparison of hydrate saturation in Problem 3, part 1
Depressurization (3_2)
The results for all the simulators match except for a small difference in the shape
of the hydrate front (Figure 7-4).
0.45
0.4
0.35 UH_3days
0.3
Tough_3days
0.25
SH
STOMP_3days
0.2
0.15 MH21_3days
0.1 HRS_3days
0.05
0
0 5 10 15 20 25 30 35
0.6
0.5
UH_5days
0.4 Tough_5days
STOMP_5days
0.3
SI
MH21_5days
0.2 HRS_5days
0.1
0
0 5 10 15 20 25 30 35
Grid Block Number
40
Problem 4
Thermal Stimulation in Radial Coordinates
Problem 4 addresses thermal stimulation of a hydrate reservoir in 1-D radial
coordinates. The radial hydrate saturation distribution at several times show that
all the plots almost fall on each when plotted against r2/t. The results from our
simulator matched with those from other simulators (not shown).
The numerical model used is a finite-volume simulator that takes into account
heat transfer, multiphase fluid flow and equilibrium thermodynamics of hydrates.
Four components (hydrate, methane, water and salt) and five phases (hydrate,
gas, aqueous-phase, ice and salt precipitate) are considered in the simulator.
Water freezing and ice melting are tracked with primary variable switch method
(PVSM) by assuming equilibrium phase transition. Equilibrium simulation
method is used here because kinetics of hydrate formation and dissociation are
relatively fast in the field-scale. This simulator has been validated against
several other simulators for the problems in the code comparison study
conducted by US DOE.
Gas
Warm water production
120 120
Aquifer
The objective of this study is to identify optimum production strategies for gas
production from Class 2 hydrate reservoirs through numerical simulation. The
domain selected as the base case is a quarter five-spot of size 120m x120m
x10m (Figure 7-6). Initial temperature and pressure are assumed to be 7.5°C
and 9MPa, respectively, which lie in the hydrate stable zone. The bottom 2m of
the domain is an aquifer layer (SA=1.0) and the top 8m is a hydrate layer with a
hydrate saturation, SH of 0.6 and aqueous saturation, SA of 0.4. There is no heat
and mass transfer though the side boundaries due to symmetry. There is only
41
heat transfer, but no mass flow through the top and bottom boundaries due to
impermeable shale layers. The effect of injection temperature, injection pressure
and production well pressure on gas and water production is studied. The
saturation histories encountered in these simulations will be modeled at the pore
scale for transport properties.
Simulations were run for different injection pressures, injection temperatures and
production pressures for 3000 days and total production of gas was compared for
the above parameters.
For the case of no injection, the dissociation is due to pressure falling below the
hydrate stable pressure due to depressurization at the production well. The heat
of dissociation comes from surroundings, decreasing the temperature of the
reservoir. Ice starts forming if the pressure goes below quadruple point pressure.
After all the hydrates dissociate, the temperature again starts rising by the heat
from surroundings.
For the case of warm water injection, the pressure of injection has to be higher
than the reservoir pressure for the hot water to go in. The temperature rise is
higher for higher temperature and higher injection pressure (injection flow rate
increases). But if injection pressure is high, the average pressure in the reservoir
increases, slowing the dissociation of hydrates (and even formation of additional
hydrates) before the warm water reaches a certain region. If production pressure
and temperature are both high, the rate of production of gas increases. The total
production of gas also depends on the production pressure, and for different
production pressure the optimum injection conditions vary.
Figure 7-7 shows total production for the production well pressure of 2MPa. The
injection temperature was kept constant at 20C and injection pressure was
varied. The results were compared against the no injection or depressurization
only case. When warm water is injected at a higher pressure but at a relatively
low temperature (20C in the present case) the gas production rate decreases
with increasing injection pressure. This is because the average pressure of the
reservoir domain increases; dissociation of hydrate slows down. In case of
5MPa of injection pressure, the total production of gas increases because water
occupies some pore space that would have been occupied by gas during
depressurization. At higher injection pressure the hydrate dissociation is not
complete in 3000 days. For low temperature water injection, only
depressurization seems to be better than warm water injection.
42
Production Pressure 2MPa
Injection Temperature 20C
3.50E+06
3.00E+06
Cumulative production STDCM
2.50E+06
NoInj
5MPa
2.00E+06
10MPa
20MPa
1.50E+06 30MPa
40MPa
1.00E+06
5.00E+05
0.00E+00
0.00E+00 5.00E+07 1.00E+08 1.50E+08 2.00E+08 2.50E+08 3.00E+08
Time (s)
Figure 7-7: Cumulative production of gas with varying injection pressure, 20°C of
injection temperature and 2MPa of production pressure
Figure 7-8 shows the cumulative production of gas when production well
pressure is kept at 4MPa and injection temperature is 80°C. The injection
pressure is varied. In this case, only depressurization is slow and does not
dissociate all the hydrates present in 3000 days. With increasing injection
pressure the gas production rate increases. With an injection water of 80°C, as
the injection pressure increases more of the reservoir gets to this high
temperature which helps hydrate dissociation.
NoInj
5MPa
10MPa
20MPa
30MPa
40MPa
Figure 7-8: Cumulative production of gas with varying injection pressure at 80°C
of injection temperature and 4MPa of production pressure.
43
If injection temperature is in medium range (50°C) then injection pressure and
production pressure play an important role. Figure 7-9 and 7-10 are plots for
2MPa and 4MPa of production pressure, respectively, at 50°C of injection
temperature with varying injection well pressures. If Injection pressure rises from
5MPa to 10MPa the production almost remains same for the case of production
pressure 2MPa but decreases drastically in the case of production pressure
4MPa. This can be attributed to higher average pressure in the reservoir
domain, which hinders hydrate dissociation. In case of injection pressure of
30MPa and 40MPa the total production and rate of production increases (Figure
7-9 and 7-10), though initial rate of production falls due to increase in average
reservoir pressure, which assists hydrate formation while temperature is still not
high. The gas production rate is non-monotonic with the increase in injection
pressure.
5MPa
10MPa
30MPa
40MPa
Figure 7-9: Cumulative gas production with varying injection pressure and 2MPa
of production pressure and 50°C of injection temperature.
44
I(5MPa)
I(10MPa)
I(30MPa)
I(40MPa)
Figure 7-10: Cumulative gas production with varying injection pressure and
4MPa of production pressure and 50°C of injection temperature.
Pore-Scale Model
We have modeled the pore-level hydrate distribution in single-phase flow (no free
gas phase) in order to estimate transport properties of hydrate bearing
sediments. A basic element of porous media is a pore throat. We have
assumed a simplified cylindrical geometry for a pore throat. We have also
assumed that this cylindrical throat is at a temperature and pressure where
hydrates can form if the methane content is high enough. Water saturated with
methane at a higher temperature flows into this pore throat. As the water passes
through this throat, temperature falls, hydrates can form at a low-enough
temperature region. Characteristic times for diffusion, heat conduction, reaction
and flow are compared.
Results from this study indicate that temperature is the coldest near the pore wall
because heat is removed through the wall. Hydrate forms near on the wall and
builds up. The rate at which hydrate layer builds up depend on the temperature
gradient, flow rate etc. These observations will be used to build a model for
hydrate deposition in a medium with distributed pore size distribution. As the
cold water passes through a collection of pores, the hydrate saturation at the
pore scale will vary from pore to pore. We will use these results to calculate
permeability of porous media as a function of overall hydrate saturation. We will
also develop models for multiphase flow during hydrate deposition.
45
CONCLUSIONS
Our in-house simulator performs as well as the other simulators in the code
comparison study. For Problem 3 with ice formation, our simulator, STOMP, and
MH21 performed better than the other simulators. For warm water injection,
production well pressure, injection temperature and pressure play an important
role in the production of gas from hydrate deposits. For high injection
temperature, the higher pressure increases the flow of warm water (heat) in the
reservoir making the production rate faster, but if injection temperature is not high
then only depressurization is the best method of production. At intermediate
injection temperature, the production rate changes non-monotonically with the
injection pressure. These parameters should be chosen carefully to optimize
recovery and recovery rate of gas. This paper addresses a very simple
homogeneous domain. Realistic reservoirs would have heterogeneity in
sediments as well as hydrate distribution, which need to be taken into account.
Models are being developed in Dr. Hirasaki’s group to address the variation in
hydrate saturation in marine sediments.
Students
Jyoti Phirani [continuing Chemical Engineering PhD Thesis, 2010] Modeling of
Gas Production from Marine Hydrate Deposits
46
Task 8: Seafloor and Borehole Stability
Approach
We have continued to expand our database on flow properties of fine-grained
sediments from oceanic hydrate settings. We have interacted and collaborated
with other hydrate researchers to make sure we have digital reprints of the
papers and reports. We have reviewed the papers and compiled data sets. We
are separating the references based on type of measurement (e.g., mechanical
property, thermal property, flow property, etc.) and how the measurement was
47
made. It is critical to understand the measurement technique as it has a large
impact on the application of the results to hydrate systems.
Based on exciting results pointing out the importance of flow and hydrate
accumulation in heterogeneous and anisotropic porous media (Task 6), we have
developed a new sample procedure for horizontal and vertical permeability
measurements. The approach uses standard flow-through permeability
techniques with existing laboratory equipment at Rice. We have adapted our
permeability chamber to accommodate vertically and horizontally oriented cores.
The correlation of permeability measurements and NMR log data (T2 relaxation
times) has allowed us to develop a permeability model for Keathley Canyon. We
based our model on the Schlumberger-Doll Research (SDR) equation relating
permeability and T2 (Kenyon et al., 1989).
48
core-measured permeability with LWD measurements of T2LM from the NMR log.
From this we have developed a lithology-dependent A based on the response of
the gamma ray log.
Figure 8.1: Measured permeability from Keathley Canyon (red squares) compared with the
NMR-derived permeability with the modified SDR equation. The modified SDR equation
allows high-resolution (<1m) estimates of permeability which can be incorporated in
numerical models of flow through hydrate systems.
Conclusion
We are nearing completion of a physical properties database that will help
advance our understanding of flow, strength, and compression in hydrate bearing
sediments. Along with this, we are working with other hydrate researchers to
establish how we can fix gaps in multi-phase flow properties in coarse-grained
materials and strength properties in fine-grained sediments with low-to-moderate
hydrate saturation.
Our permeability measurements and combination with LWD data has enhanced
our knowledge on flow and transport properties in fine-grained sediments. We
also feel it expands the use of NMR logs to get permeability in multiple
49
lithologies. Historically the SDR equation has been employed in reservoir
systems; we have adapted it for marine hydrate settings.
Approach
Our basin-scale models of fluid flow have been coupled to a slope stability
calculation that we are testing in the absence of hydrate. Preliminary tests are
able to predict when slope failures will exist under different pore pressure
regimes and seafloor slopes. In summer 2008, I plan to have Justin Stigall
(Earth Science graduate student, Task 8) work with Sayantan Chatterjee
(Chemical & Biomolecular Engineering graduate student, Task 6) to implement
the stability calculations in the two-dimensional models that couple
sedimentation, consolidation, fluid flow, and hydrate accumulation. Currently we
are looking at infinite slope analyses (Figure 8.2) which are computationally
inexpensive and give quick assessment of stability. In future tests, we will look at
more complex failure modes through more developed stability models (e.g., 2-D
method of slices) applied to zones identified as potentially unstable in infinite
slope analyses.
50
Figure 8.2: Schematic representation of slope stability calculations using the infinite slope
approximation. The model compares the forces resisting failure (Fr) to those driving failure
(Fgs) to calculate a factor of safety (FS). Failure, or slope instability, occurs when FS < 1.
The infinite slope analysis (Equation 8.3) assess stability by comparing the
gravitational forces driving failure to the frictional forces resisting failure.
FS =
[ ]
c + (Sv − ρ w gz )cos 2 α − P * tan φ
Equation 8.3
(Sv − ρ w gz )sin α cosα
Unstable condition exist when FS < 1. Sv is the overburden stress, ρw is the
density of water, g is the acceleration due to gravity, z is the depth below sea
level, α is the slope of the seafloor, P* is fluid overpressure, c is cohesion, and φ
is the angle of friction.
Conclusions
Modeling of stable and unstable conditions is complete with an infinite slope
approximation in the absence of hydrate. We need to incorporate the same
stability analyses into the hydrate models to continue working on addressing
(in)stability in hydrate settings. This study is on target.
Approach
We have worked directly with the USGS, MIT, GATech, and LBNL to integrate
our data sets and address the best technology for estimating physical properties.
This has included work I completed with the USGS to develop testing procedures
and preparation techniques for working with pressurized hydrate cores from
51
India hydrates program (NGHP). I have also worked with W. Waite (USGS
Woods Hole) to develop data processing stream and interface for working with
data from the Gas Hydrate and Sediment Testing Laboratory Instrument
(GHASTLI) such that it is available in near-real time. Last, I attended a DOE-
sponsored workshop in Atlanta, GA to assess the current techniques used for
hydrate physical properties measurements, identify the strengths in those
measurements, and identify key data gaps.
Conclusions
Collaboratively with other DOE-sponsored researchers, we are finishing a review
paper of the existing data to capture the state of the art and the status of physical
properties measurements in hydrate bearing sediments. Along with this, we
have isolated the gaps and are pursuing options to fill these gaps.
Students
H. Daigle, PhD student, Earth Science
L. Ashley Hubbard [MS Earth Science Thesis, 2008] - Investigation of fluid
relationships at diverse sites in the northern Gulf of Mexico using dissolved
ion concentrations and strontium isotopes
Steinar Hustoft [visiting Ph.D. student from University of Tromso, Norway 2008] –
Studied sediment properties and fluid flow history for the Norwegian margin
52
Hydrate Activities
Physical Properties of Hydrate Bearing Sediments participant [16-19 March
2008, Atlanta, GA]
Offshore Technology Conference participant [5-8 May 2008, Houston TX]
DOE/JIP Site Selection Working Group member [2007-present]
Geofluids Editorial Board member [2007-present]
Steering Committee Member for IODP Workshop “Addressing Geologic Hazards
through Ocean Drilling” [26-30 August 2007, Portland, OR]
Reviewer for Basin Research; Geochemistry, Geophysics, and Geosystems;
Journal of Geophysical Research; Marine and Petroleum Geology; Research
Council of Norway [2007-present]
Earth Science Overview and Rice Lab Tour for Citizens School Students [28
September 2008]
FS – factor of safety
Fr – forces resisting failure
Fgs – forces driving failure
γ – bulk unit weight of the sediment
α – angle of seafloor
φ – internal angle of friction
P – pore fluid pressure
c – cohesion
Sv – total vertical stress
P* - fluid overpressure
ρw – density of water
g – acceleration due to gravity
z – depth below sea-level
53
References
ASTM International, 2006, Standard Test Method for One-Dimensional
Consolidation Properties of Saturated Cohesive Soils Using Controlled-Strain
Loading, Standard D4186-06.
Dillon, W.P., Nealon, J.W., Taylor, M.H., Lee, M.W., Drury, R.M., Anton, C.H.,
2000, Seafloor collapse and methane venting associated with gas hydrate on
the Blake Ridge; causes and implications to seafloor stability and methane
release, in Geophysical Monograph 124, Paull, C.K. and Dillon, W.P. (eds),
211-233.
Dugan, B., 2008, Fluid Flow in the Keathley Canyon 151 Mini-Basin, Northern
Gulf of Mexico, Marine and Petroleum Geology,
doi:10.1016/j.marpetgeo.2007.12.005.
Flemings, P.B., Liu, X., Winters, W.J., 2003, Critical pressure and multiphase
flow in Blake Ridge gas hydrates, Geology, 31(12), 1057-1060.
Hornbach, M.J., Saffer, D.M., Holbrook W.S., 2004, Critically pressured free-gas
reservoirs below gas-hydrate provinces, Nature, 427, 142-144.
Kenyon, W.E., Howard, J.J., Sezinger, A., Straley, C., Matteson, A., Horkowitz,
K., Ehrlich, R., 1989, Pore-size distribution and NMR in Microporous Cherty
Sandstones, Transactions of the SPWLA 30th Annual Logging Symposium,
Denver CO (USA), 11-14 June 1989, paper LL.
Kleinberg, R.L., 1999, Nuclear magnetic resonance, In Wong, P.-Z. (ed),
Methods in the Physics of Porous Media, 337-385, San Diego: Academic
Press.
Kleinberg, R.L., Flaum, C., Straley, C., Brewer, P.G., Malby, G.E., Peltzer III,
E.T., et al., 2003, Seafloor nuclear magnetic resonance assay of methane
hydrate in sediment and rock. Journal of Geophysical Research 108(B3),
2137, doi:10.1029/2001JB000919.
Vogt, P.R., and Jung, W., 2002, Holocene mass wasting on upper non-Polar
continental slopes—due to post-Glacial ocean warming and hydrate
dissociation?, Geophysical Research Letters, 29(9), 1341,
doi:10.1029/2001GL013488.
54
Task 9: Geophysical Imaging of Gas Hydrate and Free Gas Accumulations
For this task in particular, and others in general, we have successfully initiated
collaboration with National Institute of Oceanography (NIO), India. We intend to
demonstrate geophysical imaging with multichannel seismic data from the
Krishna-Godavari (K-G) basin in the Indian east coast. NIO scientists are
tentatively scheduled to visit Rice University in July.
Though Priyank Jaiswal has not started working on the gas hydrate
project officially, he is helping NIO in preliminary processing of the selected
seismic line and thus is working towards subtask 9.1. The results of the
preliminary processing will be shown at Rice University by the NIO scientists
during their scheduled visit.
It is estimated that Priyank Jaiswal will start working on the project directly
as a Post Doc sometimes in summer.
55
Task 10: Technology Transfer Activities
The members of the research team gave presentations at the NETI
Hydrate Merit Review at Golden, Colorado on September 18-20, 2007. The
group gave a briefing to Rick Baker on April 14, 2008 at Rice University.
Publication and Presentations by Tasks
Task 5
**Snyder, G.T., A. Hiruta, R. Matsumoto, **G.R. Dickens, H. Tomaru, R.
Takeuchi, J. Komatsubara, Y. Ishida & **H. Yu, 2007. Pore water profiles and
authigenic mineralization in shallow marine sediments above the methane-
charged system on Umitaka Spur, Japan Sea. Deep-Sea Research (II), 54:
1216-1239.
56
G. Bhatnagar, W. G. Chapman, G. R. Dickens, B. Dugan, and G. J. Hirasaki,
“Sulfate-methane transition as a proxy for average methane hydrate
saturation in marine sediments,” Geophysical Research Letters, VOL. 35,
L03611, doi:10.1029/2007GL032500, 2008.
G. Bhatnager, G. J. Hirasaki, W. G. Chapman, B. Dugan, G. R. Dickens,
“Quantifying methane hydrate saturation in different geologic settings,”
The University of Texas, November 26, 2007.
G. Gu, G. Bhatnager, G. Dickens, W. Chapman, G. J. Hirasaki, poster, “Effects of
Seafloor Temperature on the Distribution of Methane Hydrate,” American
Geological Union, Fall Meeting, 10-14 December 2007, San Francisco.
G. J. Hirasaki, “Effect of Overpressure on Gas Hydrate Distribution,” poster
presentation at the 6th International Conference on Gas Hydrate,
Vancouver, BC, July 6-10, 2008. (Appendix B)
Task 7 Presentations
July 2008 ICGH, Vancouver, Canada
March 2008 Rice University Consortium on Processes in Porous Media
Sept 2007 DOE-NETL Gas Hydrate Merit Review, Golden, CO
Hydrate Presentations
March 2008 Rice University Consortium on Processes in Porous Media
Sept 2007 DOE-NETL Gas Hydrate Merit Review, Golden, CO
Aug 2007 IODP Geohazards Workshop, Portland, OR
57
Bhatnagar, G., W.G. Chapman, G.R. Dickens, B. Dugan, G.J. Hirasaki, 2008,
The sulfate-methane transition as a proxy for average methane hydrate
saturation in marine sediments, Geophysical Research Letters, 35, L03611,
doi:10.1029/2007GL032500.
Bhatnagar, G., W.G. Chapman, G.R. Dickens, B. Dugan, G.J. Hirasaki, 2007,
Generalization of gas hydrate distribution and saturation in marine sediments
by scaling of thermodynamic and transport processes, American Journal of
Science, 307(6), 861-900, doi:10.2475/06.2007.01.
Task 8 Abstracts
Daigle, H., Dugan, B., in review, Extending Nuclear Magnetic Resonance Data
for Permeability Estimation in Low-Permeability Sediments, 2008
Schlumberger Information Solutions Global Forum, Paris, France, 6-9
October 2008.
58
COST PLAN / STATUS
Phase 1 Phase 2: Year 1 (June 2007 - May 2008)
Baseline
Reporting 7/1/07 TO 9/30/07 10/1/07 TO 12/31/07 1/1/08 TO 3/31/08 4/1/08 TO 6/30/08
Quarter
Baseline Cost Plan (SF-
424A)
Federal Share $ 3,624 $ 80,003 $ 80,003 $ 80,003 $ 80,003
Non-Federal Share $ 1,004 $ 28,653 $ 28,653 $ 28,653 $ 28,653
Total Planned $ 4,628 $ 108,656 $ 108,656 $ 108,656 $ 108,656
Cumulative Baseline Cost $ 4,628 $ 113,284 $ 221,940 $ 330,596 $ 439,252
Actual Incurred Cost
Federal Share $ 3,082 $ 59,364 $ 20,610 $ 55,903 $ 93,151 *
Non-Federal Share $ 1,091 $ 18,616 $ 33,647 $ 52,085 $ 13,797 *
Total Incurred $ 4,173 $ 77,980 $ 54,257 $ 107,988 $ 106,948 *
Cumulative Costs $ 4,173 $ 82,153 $ 136,410 $ 244,398 $ 351,346 *
Variance (plan-actual)
Federal Share $ 542 $ 20,639 $ 59,393 $ 24,100 $ (13,148)*
Non-Federal Share $ (87) $ 10,037 $ (4,994) $ (23,432) $ 14,856 *
Total Variance $ 455 $ 30,676 $ 54,399 $ 668 $ 1,708 *
Cumulative Variance $ 455 $ 31,131 $ 85,530 $ 86,198 $ 87,906 *
* These quantities are for the period April-May, 2008. Financial data for June
has not yet been posted.
The cumulative variance at the end of the period has the appearance
of the project being under spent. The financial data does not include the June
1-19 period. We expect by June 19, 2008 the cumulative variance will be
close to zero. This period will include faculty summer salaries and a
correction for the stipend of a graduate student (Gaurav Bhatnagar) who was
paid from another account last fall.
59
Milestone Plan/Status
60
The recipient shall summarize, quantitatively,
the conditions for the absence, presence, and
distribution of gas hydrates and free gas in 1-D
systems by expressing the conditions in terms
of dimensionless groups that combine
thermodynamic, biological and lithologic
transformation, and transport parameters.
6.3 Compositional effect on BSR 7/07 12/08
The recipient shall add to the numerical model,
developed under this task, a chloride balance
and multi-hydrocarbon capability specifically to
investigate how hydrocarbon fractionation
might affect Bottom Simulating Reflectors
(BSRs).
6.4: Amplitude Attenuation and chaotic zones 3/09
due to hydrate distribution
The recipient shall simulate preferential
formation of gas hydrate in coarse-grained,
porous sediment in 2-D by linking fluid flux to
the permeability distribution.
6.5: Processes leading to overpressure 3/08 done
The recipient shall quantify, by simulation and
summarize by combination of responsible
dimensionless groups, the conditions leading
to overpressure to the point of sediment
failure.
6.6 Concentrated hydrate and free gas 3/08 started
The recipient shall, using 2-D and 3-D models,
simulate lateral migration and concentration of
gas hydrate and free gas in structural and
stratigraphic traps.
6.7 Focused free gas, heat and salinity 9/09
The recipient shall quantify, using 2-D and 3-D
model simulations and comparisons to
available observations, the factors controlling
the process of localized upward migration of
free gas along faults and lateral transfer to
dipping strata that can lead to chaotic zones
and possible accumulations of concentrated
hydrate.
6.8 Sulfate profile as indicator of methane flux 7/07 done
61
The recipient shall compute, for systems
where data on the sulfate profile is available,
the oxidation of methane by sulfate and shall
indicate the perceived level of effect on gas
hydrate accumulation and the data’s value as
an indicator of methane flux.
6.9 Application of models to interpretation of 6/10 6/10
case studies.
The models developed in Task 6 will be
applied to case studies in the interpretation of
each of the other tasks.
7. Analysis of 7.1a Pore scale model development and 1/08 6/08
production Hydrate code comparison
Code
strategy For this milestone, we will develop pore-scale compa
models of hydrate accumulation by simulation. rison is
Our hydrate code will be used to solve a set of done.
problems formulated by the Code Comparison
Study group. Our results will be compared with
those of other hydrate codes.
Should be changed to: 6/08
Reason: The starting date was moved to 6/07
Status: Code comparison study is 80%
complete.
7.1b Petrophysical and thermophysical 1/09 6/09
properties of hydrate sediments from pore-
scale model
For this milestone, we will assume the pore-
scale models of hydrate accumulation
developed in the last milestone and estimate
transport properties as a function of hydrate
and gas saturations.
Should be changed to: 6/09
Reason: The starting date was moved to 6/07
Status: Have not started
7.2a Modeling of several production strategies 1/10 6/10
to recover gas from marine hydrates
Several production strategies would be
modelled using the transport property
correlations developed in the previous
milestone. Optimal strategies will be identified.
Should be changed to: 6/10
62
Reason: The starting date was moved to 6/07
Status: Have not started
7.2b Effect of marine reservoir heterogeneities 12/10 6/10
on production of methane
Reservoir heterogeneity anticipated in marine
environments (known or determined through
other tasks) would be incorporated.
Appropriate hydrate distributions, either
constrained from experimental data or
mechanistic simulations (Task 5) would be
used. Sensitivity of gas production to the
heterogeneities would be calculated.
Should be changed to: 6/11
Reason: The starting date was moved to 6/07
Status: Have not started
8. Seafloor 8.1a Collection of data 05/08 Compl
and borehole eted
We have collected the published data and are
stability working it into a data base. We are also
working on a review paper summarizing the
state of the art settings. This will include
laboratory experiments, field data, published
results, and unpublished data.
8.1c Complete database 10/09 On
target
We are organizing the data from task 8.1a into
a format that can be searched and used by
researchers trying to understand mechanical
behavior of hydrate-bearing sediment. We will
also identify key gaps in the database for
focusing future hydrate research endeavors.
We have started exchanging these data with
the modeling components of this project.
8.2a Link database with models 08/08 On
target
We have started passing data along to the
modeling groups so they can use sediment
properties from hydrate provinces as they
simulate hydrate accumulation and production.
8.2b Add sediment stability to models 10/08 On
Standard stability calculations have been target
implemented in a standard basin model. Now
that it is functional we will work with the
hydrate accumulation model to add a stability
calculation to the 2-D models.
63
8.2c Conditions for (in)stability 9/09 On
After implementing the stability model in the target
hydrate accumulation code, we can explore
the conditions (e.g., hydrate dissociation, sea-
level fall) that could drive slope failure and
hydrate/methane release or lead to borehole
failures during production.
9 9.1 Preliminary processing and inversion of 8/08
Geophysical seismic data.
imaging of Perform conventional seismic reflection
hydrate and processing, velocity analysis, travel time
free gas tomography, and other analyses as deemed
appropriate and necessary.
9.2: Final 1-D elastic and 2-D acoustic 8/09
waveform inversion.
Apply 1-D elastic and 2D acoustic inversions
on data obtained from subtask 9.1 to derive
determine high-resolution elastic and acoustic
properties.
9.3: Rock physics modeling. 8/10
Apply rock physics models to the developed
seismic models to estimate hydrate saturation
and lithology through application of well log
data in conjunction with data from subtask 9.2.
For this subtask we shall seek to collaborate
with research being conducted under
separately funded DOE-NETL projects (DE-
FC26-05NT42663 with Stanford University,
"Seismic-Scale Rock Physics of Methane
Hydrate" and others as applicable).
64
RICE UNIVERSITY
A T HESIS S UBMITTED
IN PARTIAL F ULFILLMENT OF THE
R EQUIREMENTS FOR THE D EGREE
Doctor of Philosophy
Clarence A. Miller
Louis Calder Professor in Chemical and
Biomolecular Engineering
Brandon Dugan
Assistant Professor in Earth Science
Houston, Texas
February, 2008
ABSTRACT
by
Gaurav Bhatnagar
Generalized numerical models for simulating gas hydrate and free gas accumu-
lation in marine sediments have been developed. These models include several
ent modes of gas hydrate and free gas occurrence, e.g., no hydrate and hydrate
with or without underlying free gas. Scaling schemes developed for these systems
help in summarizing average gas hydrate saturation for hundreds of simulation re-
sults into two simple contour plots. The utility of these contour plots in predicting
tings.
through numerical modeling that scaled depth to the SMT correlates with the aver-
age gas hydrate flux through the gas hydrate stability zone (GHSZ). Later, analyt-
as the complete gas hydrate saturation profile from the SMT depth. Application of
this theory to several sites along Cascadia Margin indicates that SMT depth can
be used as a fast and inexpensive proxy to get a first-order estimate of gas hydrate
tems is shown to have an important effect on gas hydrate and free gas saturations.
and free gas in the system, in addition to extending the base of the hydrate stabil-
ity zone below the seafloor by a relatively small depth. We also study the role of
upward free gas migration in producing long, connected free gas columns beneath
the gas hydrate layer. Finally, two-dimensional models are developed to study
the effect of heterogeneities on gas hydrate and free gas distribution. Simulation
results show that hydrate as well as free gas accumulates in relatively high satu-
dipping sand layers, and combinations of both, due to focused fluid flow.
Acknowledgements
Dr. George J. Hirasaki for being my thesis advisor and a great mentor. His
towards the completion of this thesis. In addition, his work ethics and dedication
and very interesting multidisciplinary project. I will be forever grateful to him for all
Dr. Walter G. Chapman for being my co-advisor and for his guidance and sup-
Dr. Brandon Dugan for serving on my thesis committee and always providing
valuable suggestions and comments on my thesis work. I learnt a lot about hydro-
geology from him through numerous discussions and paper revisions, and thank
Dr. Gerald R. Dickens for teaching me the subtleties of marine geology, and
earth science in general. His ideas and suggestions helped me appreciate re-
search problems from a very different perspective. I also thank him for several
stimulating discussions on gas hydrate systems and for greatly improving my tech-
I thank the Shell Center for Sustainability, the Kobayashi Graduate Fellowship,
I also thank all my friends and colleagues at Rice University for their help and
Lastly, I would like to acknowledge my parents and family in India for their con-
tinued love and support. This thesis would not have been made possible without
Gaurav Bhatnagar
Contents
Abstract ii
Acknowledgements iv
1 Introduction 1
1.1 Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.3.1 Non-dimensionalization . . . . . . . . . . . . . . . . . . . . . 33
4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
D 39
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.7 Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.2 Mathematical Model for Gas Hydrate Accumulation and AOM . . . . 118
Systems 193
Bibliography 245
A Appendix 257
A.3 Deriving Conditions for Maximum L̃s and Minimum L̃h . . . . . . . . 261
Glossary
1.1 The three primary unit crystal structures of gas hydrates: Structure
I, Structure II and Structure H. Pure methane and ethane form
Structure I hydrate, but certain compositions of their mixtures
undergo phase transition to form Structure II hydrates. The notation
512 62 indicates that a particular water cage consists of 12
pentagonal faces and 2 hexagonal faces [Taken from Sloan (2003)]. 2
5.5 Steady state dissolved methane concentration profiles for the case
of methane supplied from deeper sources and without the effect of
sedimentation (Pe1 = 0). As Pe2 increases in magnitude (that is,
upward external flux becomes more important), the gas hydrate
and free gas layers approach each other and co-exist at the base of
the GHSZ only if the critical flux Pe2 = −5 is exceeded. Parameters
used for these simulations are: Pe1 = 0, c̃wm,ext = 1.0, Pe2 varies for
the four plots. Base-case seafloor parameters were used for this
simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.6 Parameter space of Pe1 and β showing three different fields of gas
hydrate distribution. As β increases, with Pe1 held constant, the
system moves from the state of no gas hydrate to gas hydrate
without free gas below and finally to gas hydrate with free gas below. 64
5.8 Collapse of the curves shown in Figure 5.7 into one single pair of
curves achieved by rescaling and plotting them in a manner such
that the x-axis represents the total methane generated within the
GHSZ (equation 5.38). Base-case seafloor parameters were used
for this simulation and Pe2 = 0. . . . . . . . . . . . . . . . . . . . . . 68
5.11 Average gas hydrate saturation (equation 5.38) contours plotted for
the case of biogenic methane generated in situ (Pe2 = 0). The set
of curves plotted in Figure 5.8 are also shown on the left side of the
diagram. Base-case seafloor parameters were used for this
simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.12 Contours of gas hydrate flux Pe1 hSh i plotted along with the curves
separating the two regions of gas hydrate occurrence in Figure
5.10 for the case of non-zero sedimentation and |Pe1 | < |Pe2 |. Gas
hydrate saturation can be calculated by dividing the contour values
with . For example, if Pe1 = 1, these contours directly represent the
average gas hydrate saturation in the pore space. Base-case
seafloor parameters were used for this simulation. . . . . . . . . . . 74
xvii
6.1 Contours of gas hydrate flux Pe1 hSh i plotted along with the curves
separating the two regions of gas hydrate occurrence, for the case
of non-zero sedimentation and |Pe1 | < |Pe2 |. Gas hydrate saturation
can be calculated by dividing the contour values with . For
example, if Pe1 = 1, these contours directly represent the average
gas hydrate saturation in the pore space . . . . . . . . . . . . . . . . 85
6.3 Gas hydrate saturation contours averaged over the entire GHSZ for
systems where all methane is furnished through in situ biogenic
reactions. Diffusive losses dominate at low values of Peclet number
(Pe1 ), implying that methane generation within the GHSZ has to
increase to form any gas hydrate. The set of dashed curves,
marked (a) and (b), represent the intermediate region of gas
hydrate formation without free gas, with the region of no gas
hydrate formation for x-axis values lesser than for curve (a) and gas
hydrate with free gas immediately below for x-axis values greater
than those for curve (b) (Bhatnagar et al., 2007). Average gas
hydrate saturation at different gas hydrate settings can be obtained
from this single contour map. . . . . . . . . . . . . . . . . . . . . . . 87
xviii
6.5 Steady state heterogeneous gas hydrate (solid curves) and free
gas (thick dashed curves) distribution at Sites 994, 995 and 997
(Blake Ridge). The biogenic parameters remain constant for both
cases. Pe1 = 0.05 and β = 2.25 listed in the insets correspond to an
average sedimentation rate of 11 cm/kyr and TOC content of 1.5 %
(with 40 % available organic carbon), respectively. Other seafloor
parameters and constants are the same as those listed in Table 6.1
for Blake Ridge Site 997. (A) Site 994, simulated only with a
biogenic in situ methane input, shows an isolated gas hydrate layer
that does not extend to the base of the GHSZ. Free gas starts at
550 mbsf, leaving a 90 m zone beneath the GHSZ devoid of any
free gas, thus explaining the absence of any BSR at this site. (B)
Using the same biogenic methane input as in case (A), but with a
low upward flux recycling methane back into the GHSZ, gas
hydrate as well as free gas extend to the base of the GHSZ at Sites
995 and 997, resulting in BSRs at both these locations. . . . . . . . 99
6.8 Effect of variable upward fluid flux Pe2 on gas hydrate saturation
with constant biogenic in situ methane generation. Pe1 , Da and β
were set to 0.1, 5.0 and 3.0,respectively, for all three cases.
Keeping biogenic input constant and increasing the magnitude of
net methane flux Pe1 + Pe2 (by increasing Pe2 ) causes higher
steady state gas hydrate saturations as well as shallower depths to
the top of the hydrate layer. . . . . . . . . . . . . . . . . . . . . . . . 113
7.2 Effect of Damkohler number (DaAOM ) and Peclet numbers (Pe1 , Pe2 )
on steady state profiles. Pe1 + Pe2 = −10 for all simulations (Note
different y-axis scale for each plot). (a) Sulfate and methane
profiles for DaAOM = 108 (solid curves) and DaAOM = 106 (dashed
curves). The hatched regions compare the thickness of the SMT
zone for the two cases. (b) Simulations for different sets of Pe1 and
Pe2 , but with Pe1 + Pe2 = −10. Overlap of methane and sulfate
profiles shows that the sum Pe1 + Pe2 controls the concentrations.
(c) Gas hydrate saturation profiles do not depend only on Pe1 + Pe2 ,
because different Pe1 causes distinct residence times of gas
hydrate within the GHSZ. . . . . . . . . . . . . . . . . . . . . . . . . 123
xx
7.4 Relationship between Pe1 hSh i and scaled SMT depth (L̃s = Ls /Lt )
for several seafloor depths. Points corresponding to different
Cascadia Margin sites are plotted to show how AGHS can be
estimated from L̃s using this plot (Table 7.1). . . . . . . . . . . . . . 126
8.2 The function g(z̃), equation(8.13), plotted for three different porosity
parameters. The dashed line represents the case of zero
compaction η = 0, the dash-dot curve represents the case η = 2/9
(or φ0 = 0.3), while the solid curve denotes the case η = 6/9 (or
φ0 = 0.7). The parameter γ = 9 (or φ∞ = 0.1) for all cases. . . . . . . 139
8.8 Relationship between average gas hydrate flux (Pe1 hSh i) and
scaled SMT depth (L̃s ) for different seafloor depths. Curves
representing analytical solutions are also compared with steady
state numerical simulations of Bhatnagar et al. (2008). Shallow
SMT depths indicate higher methane flux from deeper sources
causing higher average gas hydrate flux (and saturations) through
the GHSZ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
9.3 Pressure profiles shown in Figure 9.2 plotted together for all four
cases. Lithostatic stress profiles (dashed curves) and the
corresponding pore pressure profiles (solid curves) for the same
value of Nsc are color-coded together. . . . . . . . . . . . . . . . . . 186
10.3 Gas hydrate and free gas saturation profiles at time t̃ = 0.7 for
Nsc = 1000. Gas starts to migrate to the base of the GHSZ when it
exceeds the critical saturation. . . . . . . . . . . . . . . . . . . . . . 208
10.5 Gas hydrate and free gas saturation profiles at time t̃ = 1 for
Nsc = 1000. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
10.7 Gas hydrate and free gas saturation profiles at the critical-state for
Nsc = 1000. A deep connected free gas column forms beneath the
GHSZ and is trapped by the high saturation hydrate layer. . . . . . . 212
10.8 2-D contour plot of hydrate and free gas saturations at the
critical-state for Nsc = 1000. . . . . . . . . . . . . . . . . . . . . . . . 212
10.10Gas hydrate and free gas saturation profiles at the critical-state for
Nsc = 10 (t̃ ∼ 1.5). Compared to the thick gas column in Figure
10.7, only a relatively thin free gas column forms beneath the GHSZ
in this case before gas pressure becomes equal to the overburden. . 215
10.13Free gas saturation contours at time t̃ = 0.6 for the vertical fracture
system shown schematically in Figure 10.11. Dashed lines denote
the position of the fracture. . . . . . . . . . . . . . . . . . . . . . . . 219
10.15Free gas saturation contours at time t̃ = 1.0 for the vertical fracture
system shown schematically in Figure 10.11. Peak gas saturation
is about 50% within the fracture, while lateral gas migration from
the fracture causes gas saturation in neighboring gridblocks to
increase to about 30%. . . . . . . . . . . . . . . . . . . . . . . . . . 220
10.17Free gas saturation contours at time t̃ = 1.5 for the vertical fracture
system shown schematically in Figure 10.11. Peak gas saturation
is 67% within the fracture, while lateral gas migration from the
fracture causes gas saturation in neighboring gridblocks to increase
to about 40%. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
xxvi
10.19Gas hydrate and free gas saturation at time t̃ = 2.25. Since the high
permeability sand layer extends uniformly across the lateral section,
hydrate and free gas saturation are also laterally homogeneous. . . 225
10.21Gas hydrate saturation contours at t̃ = 0.25 for the low dip angle
with ”true” axis scale, i.e., no vertical exaggeration. . . . . . . . . . . 228
10.24Gas hydrate and free gas saturation contours at t̃ = 0.3 for the case
of higher dip angle and with a vertical exaggeration (VE) of about
2:1. Color axis for the hydrate saturation contour plot is scaled to a
maximum of 15%. Dashed lines denote the position of the sand
layer within the low permeability sediment matrix. . . . . . . . . . . . 232
xxvii
10.25Gas hydrate and free gas saturation contours at t̃ = 0.25 for the
case of combined fracture and lower dip angle sand layer, with a
vertical exaggeration (VE) of about 2:1. Color axis for the hydrate
saturation contour plot is scaled to a maximum of 15%. Dashed
lines denote the position of the fracture and the sand layer within
the low permeability sediment matrix. . . . . . . . . . . . . . . . . . 234
10.26Gas hydrate and free gas saturation contours at t̃ = 0.75 for the
case of combined fracture and lower dip angle sand layer, with a
vertical exaggeration (VE) of about 2:1. Dashed lines denote the
position of the fracture and the sand layer within the low
permeability sediment matrix. . . . . . . . . . . . . . . . . . . . . . . 235
Tables
Chapter 1
Introduction
Gas hydrates are inclusion compounds formed by the encapsulation of low molecu-
lar weight gases (guest molecule) in hydrogen bonded cages formed by water (host
molecule) (Figure 1.1) (Kvenvolden, 1993; Sloan and Koh, 2007). This clathrate
structure is stabilized by weak van der Waals interactions between the guest and
host molecules. Depending on the gas (guest) composition, hydrates can form
one of three primary crystal structures: Structure I (sI), Structure II (sII) or Struc-
ture H (sH) (Figure 1.1). However, natural gas hydrates have methane as their
main guest component which causes them to occur mostly as sI hydrates. Small
amounts of heavier hydrocarbons such as ethane and propane or other gases like
carbon dioxide and hydrogen sulfide have also been found to be present as hy-
drate formers (Kastner et al., 1998; Milkov and Sassen, 2000; Milkov et al., 2005).
Mixtures of different hydrate formers, for example methane and ethane, can result
atures and low salt concentrations (Sloan and Koh, 2007). Gas hydrates, thus,
and permafrost regions, often accompanied by an underlying free gas layer (Kven-
2
Figure 1.1: The three primary unit crystal structures of gas hydrates: Structure I,
Structure II and Structure H. Pure methane and ethane form Structure I hydrate,
but certain compositions of their mixtures undergo phase transition to form Struc-
ture II hydrates. The notation 512 62 indicates that a particular water cage consists
of 12 pentagonal faces and 2 hexagonal faces [Taken from Sloan (2003)].
3
volden, 1988, 1993). In addition to temperature, pressure and salinity, gas hydrate
stability in marine settings is also influenced by gas composition and sediment pore
the associated free gas, is enormous and various global estimates have been
reported in the literature (Kvenvolden, 1993; Dickens, 2001; Buffett and Archer,
2004; Milkov, 2004; Klauda and Sandler, 2005). However, due to their sensitive
slumps and submarine slides (Borowski and Paull, 1997; Sultan et al., 2004). Hy-
drates have also been believed to affect past climate change. During the Late
Paleocene Thermal Maximum (LPTM), isotopic records indicate that ocean bot-
tom water temperature increased by more than 4◦ C within a short time interval (<
inorganic carbon reservoir (Dickens et al., 1995; Katz et al., 1999; Kennett et al.,
et al., 1995).
Thus, understanding the factors responsible for the accumulation and distribu-
climate change. The overall objective of this thesis is to develop general numerical
models to characterize the spatial and temporal distribution of gas hydrates in sedi-
4
ments due to different sources of methane. Apart from developing these numerical
models, the different modes of gas hydrate occurrence are also characterized on
the basis of a few dimensionless groups. This makes our model results applicable
rently reported in the literature. A new method for directly evaluating gas hydrate
distribution from pore water sulfate profile as the main proxy is also developed.
This method can be used to predict gas hydrate saturation in a relatively fast and
1.1 Organization
This thesis is divided into twelve chapters. Chapter 2 provides a brief background
of natural gas hydrate systems and motivation to study them in a generalized man-
ner. Current state of the art models of marine gas hydrate systems and their short-
comings are also reviewed. Chapter 3 discusses the phase equilibrium of fresh
and its role in driving fluid flow in the sediments. Chapter 5 describes various
less numerical model to various sites and compare our predictions with inferences
from proxy data at these sites. Chapter 7 develops a numerical model to estimate
proxy. In Chapter 8 we develop analytical theory and expressions relating gas hy-
Cascadia Margin sites. Chapter 9 discusses the effect of overpressure on gas hy-
drate and free gas saturations due to fast sedimentation and/or low permeability of
ulating gas hydrate and free gas (mobile) accumulation. Chapter 11 summarizes
the conclusions of this thesis and discusses future research directions arising from
this study.
6
Chapter 2
2.1 Motivation
Natural gas hydrates are known to occur worldwide in sediments on offshore con-
lation of large hydrate deposits (Kvenvolden, 1993; Sloan and Koh, 2007). This
thesis focuses only on marine gas hydrate systems, because they are estimated
Current interest in natural gas hydrates stems from three main areas:
Potential energy source: It is well known that huge amounts of methane hy-
Archer, 2004), 500-2,500 Gt (Milkov, 2004) and 74,400 Gt (Klauda and San-
dler, 2005). Such figures make methane hydrates a very promising energy
marine sites explored to date contain gas hydrate dispersed in relatively low
7
such gas hydrate systems has also been an area of active research (Moridis,
Deep water geohazard: Natural gas hydrates concentrated along continental mar-
ciate large amounts of hydrate and possibly cause sediment failure. Towards
this end, research has been done to understand how much free gas can ac-
triggering slope failures has also been recently studied (Xu et al., 2001; Sul-
tan et al., 2004). These models are relatively simple and, in general, do not
Global climate change: Rapid changes in past climate have been attributed to
carbon influx from methane hydrate dissociation. Dickens et al. (1995, 1997)
hypothesized that huge quantities of methane were released during the LPTM,
the heavier carbon isotope within a relatively short time interval (< 104 years).
Kennett et al. (2003) recently proposed the clathrate gun hypothesis, which
relates climate change during the Late Quaternary to global warming caused
(Dickens, 2003), with large carbon fluxes to and from the ocean, and could
affect present day climate if perturbed over large time periods. Studies of gas
lower temperatures (Sloan and Koh, 2007). However, both pressure and temper-
ature usually increase as a function of depth below the seafloor due to increasing
overburden and geothermal heat, respectively. This implies that the temperature
at some sub-bottom depth below the seafloor will exceed the three-phase stabil-
ity limit, thereby constraining the actual zone of hydrate occurrence to some finite
region below the seafloor (Figure 2.1). This finite zone is known as the gas hy-
drate stability zone (GHSZ) and its emergence from the three-phase equilibrium
curve is explained in detail in Chapter 3. The depth at which the geothermal gra-
dient causes the temperature to rise above the stability limit marks the base of the
Submarine gas hydrates are often found with a free gas layer below the BHSZ.
This free gas can be produced by the dissociation of hydrate (due to burial with the
sediment) beyond the GHSZ over geologic time scales (Davie and Buffett, 2001).
Alternatively, free gas can also be supplied from depth and can get sealed by
9
Figure 2.1: Schematic representation of temperature profile through the ocean and
marine sediments illustrating that hydrates are stable at depths where temperature
is less than T3 (P). The curve T3 (P) denotes the three-phase hydrate-water-gas
stability boundary [Taken from Buffett and Archer (2004)].
the low permeability hydrate layer at the BHSZ (Xu and Ruppel, 1999; Liu and
Flemings, 2007). Free gas can also migrate upwards back into the GHSZ to further
concentrate the hydrate layer. Natural gas hydrates are often identified by the
presence of a strong seismic reflection that parallels the seafloor, known as the
Results of the Ocean Drilling Program (ODP), particularly Legs 164 and 204,
have greatly enhanced our understanding of gas hydrate distribution along active
and passive margins (Paull et al., 1996; Tréhu et al., 2003). Unquestionably, at
10
many locations, lithology dictates gas hydrate distribution at the local scale (Krae-
mer et al., 2000; Weinberger et al., 2005). More important, however, are regional-
scale variations in gas hydrate distribution between and within different geologic
settings. For example, outer Blake Ridge (Leg 164) offshore South Carolina (USA)
has modest quantities of gas hydrate dispersed over immense volumes of sed-
iment, but locations with and without free gas beneath the GHSZ (Paull et al.,
2000). By contrast, Hydrate Ridge (Leg 204) on the Cascadia Margin offshore
Oregon (USA) has widely varying gas hydrate contents distributed across small
Marine gas hydrates are components of complex systems with dynamic inputs
and outputs of methane over time (Dickens, 2003). Excepting at select sites (for
example, some seeps in the Gulf of Mexico), most gas within marine gas hydrates
is methane (Kvenvolden, 1993; Milkov et al., 2005). However, this methane may
derive from two general reactions, and gas hydrates may form through two gen-
low burial depths, methanogenic archaea form biogenic methane; at relatively high
et al., 1986). Gas hydrates have been recovered from marine sediment with con-
volden, 1993; Milkov et al., 2005). More crucially, biogenic methane can be sup-
plied within the GHSZ during burial of organic carbon (in situ methane) or to the
GHSZ through upward migration of methane bearing fluids (deep methane). The
11
first input appears to dominate some gas hydrate systems, such as on outer Blake
Ridge (Paull et al., 1994; Egeberg and Dickens, 1999), while the second appears
to dominate other systems, such as along Cascadia Margin (Hyndman and Davis,
1992; Haeckel et al., 2004; Torres et al., 2004). Outputs of methane from gas
water sulfate (Xu and Ruppel, 1999; Borowski et al., 1996). Given the known varia-
tions in natural gas hydrate systems, an important issue is whether relatively simple
and straightforward numerical models can explain general features of all systems.
In particular, can changes in a few basic factors encompass a wide range of ob-
servations?
Previous numerical models for the accumulation of gas hydrate in marine sed-
iment have focused mainly on simulating the distribution of gas hydrate at specific
sites using parameters relevant to those locations. Blake Ridge (Egeberg and Dick-
ens, 1999; Davie and Buffett, 2001, 2003a; Gering, 2003) and Hydrate Ridge (Luff
and Wallman, 2003; Haeckel et al., 2004; Torres et al., 2004; Liu and Flemings,
2006) have been the subject of most studies. These models yield hydrate profiles
fairly consistent with proxy evidence, but give little insight as to how changes in
basic parameters alter the dynamics and distribution of gas hydrate. Studying ef-
new simulations to be performed for the perturbed parameter around its base case
value. Thus, separate hydrate provinces are studied as isolated examples without
hydrate distribution at the local scale necessitates modeling that accounts for de-
12
tailed geology and structure, probably in two or three spatial dimensions. However,
derstanding for how various processes affect natural gas hydrate systems.
We first develop a simple 1-D numerical model for gas hydrate accumulation
in marine sediments over geologic time-scales. Using this model, we show that
classified and linked from a mechanistic perspective. The model is later extended
tings.
The amount and distribution of gas hydrate within the GHSZ at a given location can
be quantified by several indirect techniques (Paull et al., 2000; Tréhu et al., 2004).
sediment thermal anomalies, well-log velocity and resistivity, NMR and density-
porosity logs, and seismic profiles. In concert, such methods can provide a current
”snapshot” for the presence and abundance of gas hydrate (Paull et al., 2000; Tréhu
et al., 2004). However, they give no insight as to how the gas hydrate formed or
The other main approach is to study these systems using coupled transport
models, thereby explicitly accounting for the different methane sources. Rempel
13
and Buffett (1997) modeled hydrate formation as a Stefan problem and obtained
analytical solutions for hydrate growth in a porous half-space that is cooled on its
boundary. Xu and Ruppel (1999) extended the model by including coupled mass,
momentum and energy equations to derive analytical solutions for hydrate distri-
bution in marine sediments. Based on their results, they defined a critical methane
flux required for the hydrate layer to extend to the BHSZ. However, methane was
only supplied in their model from depth, i.e. sedimentation and biogenic in situ
generation to methane were not considered. Davie and Buffett (2001) proposed a
1-D numerical model for hydrate accumulation due to biogenic sources of methane.
They used the model to predict hydrate distribution in the Blake Ridge sediments
as a case study and constrained the simulation parameters using pore water chlo-
ride anomalies from core data. Pore water sulfate gradients were used to further
constrain the different sources of methane in the Blake Ridge hydrates (Davie and
Buffett, 2003a). Gering (2003) focused on the effect of sediment compaction rates
on hydrate accumulations.
A basic problem with the aforementioned models is that they do not incorporate
both sources of methane (i.e., biogenic in situ production or methane from depth)
in a systematic manner. This causes them to be valid only at specific hydrate lo-
cations, e.g. Blake Ridge (Davie and Buffett, 2003a) or Cascadia Margin (Haeckel
et al., 2004; Torres et al., 2004). Moreover, most of these transport models use
a first-order rate law to model the kinetics of hydrate formation, the driving force
being the difference between pore water methane concentration and the local solu-
enforced by arbitrarily choosing a large value of the rate constant, with no phys-
ical basis. Porosity loss and compaction driven fluid flow is also modeled using
tary basins. Our approach aims to give a more general understanding of hydrate
patterns in sediments. None of the current models have the capability to incor-
non-empirical approach. Effects such as focused fluid flow or lateral gas migration
can also result in heterogeneous hydrate accumulations and will need at least two
Our model differs from existing models in using component balances so that the
the use of a kinetic model for hydrate formation. Further, we systematically incor-
porate both local and deeper sources of methane. The main difference, however,
lies in the way we non-dimensionalize our system, which leads to the characteri-
zation of different hydrate distributions using a few dimensionless groups. This en-
just two contour plots, which are valid for a wide range of model parameters. We
also include the effect of salinity on gas hydrate distributions so they are pertinent
to marine environments.
15
Chapter 3
3.1 Introduction
Gas hydrates can coexist with aqueous and vapor phases at specific conditions
tures lower than and pressures higher than the three-phase equilibrium curve, gas
the GHSZ to some finite sub-bottom depth (Figure 2.1). In this region, gas hydrates
can coexist with the aqueous phase; we call this region as the L-H zone. Below
this region, free gas can coexist with the aqueous phase; we call this the L-V zone.
The thickness of this GHSZ is primarily governed by the seafloor depth, seafloor
temperature, geothermal gradient and pore water salinity (Dickens et al., 1997).
Capillary inhibition due to fine-grained sediments (Clennell et al., 1999) and gas
composition have lesser impact on the GHSZ thickness. Because methane domi-
nates naturally occurring gas hydrates, we focus on the phase diagram and mod-
The three-phase equilibrium curve of gas hydrates is predicted using the statistical
thermodynamic model of van der Waals and Platteeuw (1959). The condition for
three phase stability comes from the equality of the chemical potential of water in
µH L β H β L
w = µw =⇒ µw − µw = µw − µw (3.1)
where µH L β
w , µw and µw are the chemical potentials of water in the hydrate phase,
the aqueous phase and a reference hypothetical empty gas hydrate lattice, re-
spectively. The difference in chemical potential between the reference state and
the liquid phase can be calculated using a simplified relationship first proposed by
Z T Z P
µβw − µLw L,0
∆µw ∆hw ∆νw
= − 2
dT + dP − ln(γw xw ) (3.2)
RT RT0 T0 RT 0 RT
where ∆µL,0
w is an experimentally determined chemical potential difference be-
tween the empty reference state and pure water at the reference temperature (T0 )
and zero absolute pressure, ∆hw and ∆νw are the enthalpy difference and volume
difference between the empty hydrate lattice and pure water, respectively, γw is the
activity coefficient of water, and xw is the mole fraction of water in the aqueous
phase. Due to the low solubilities of methane in water, the activity coefficient γw is
assumed to be unity, unless salinity is included. The equation of state for methane
17
proposed by Duan et al. (1992) is used to find the solubility of methane in water.
Z T
∆hw = ∆h0w + ∆Cp,w dT (3.3)
T0
where ∆Cpw is the heat capacity difference between the empty hydrate reference
et al., 1980):
0
∆Cp,w = ∆Cp,w (T0 ) + b (T − T0 ) (3.4)
The difference in chemical potential between the reference hydrate lattice and
2
µβw − µH
w
X
=− vi ln(1 − θi ) (3.5)
RT i=1
where the summation is done over both cages of structure I hydrate, vi is the ratio
of the number of type i cavities to the number of water molecules in the hydrate
structure and θi is the fractional occupancy of type i cavity by the gas molecules.
where fCH4 is the fugacity of methane in the gas phase, calculated using the Peng-
the Langmuir constant of adsorption for the small and large cages of structure I
potentials between the guest and host molecules over the cell volume, which after
" #
4π Z Ri −a 2 −wi (r)
Ci (T ) = r exp dr (3.7)
kT 0 kT
where k is Boltzmann’s constant, Ri is the cell radius, a is the core radius of the
The cell potential term wi (r) can be approximated by a Kihara potential, which
sions relating wi (r) to the Kihara potential and different cell parameters are stan-
dard (Parrish and Prausnitz, 1972; Sloan and Koh, 2007). The one-dimensional
integral in equation (3.7) can be evaluated numerically but the integrand becomes
singular at both the upper and lower integration limits. To avoid this problem we
use the Gauss-Kronrod quadrature rule (Piessens et al., 1983), which uses an
these constants at any given temperature, the equality of chemical potentials can
Z T Z P
∆µL,0
w ∆hw ∆νw
g(P ) = − 2
dT + dP − ln(γw xw ) +
RT0 T0 RT 0 RT
2
à !
X Ci fCH4
vi ln 1 − =0 (3.8)
i=1 1 + Ci fCH4
19
The parameters used in this formulation are listed in Table 3.1. The root of this
function g(P ) is the three-phase equilibrium pressure and it is evaluated using the
water (γw xw ) in solution and this effect can be incorporated in the model. We
use the Pitzer equations (Pitzer and Mayorga, 1973) to calculate water activity
∆µL,0
w 1236 J/mol 1
0
∆Cp,w -38.12 J/mol K 1
b 0.141 J/mol K2 1
T0 273.15 K 1
The three-phase equilibrium curve obtained from the solution of equation (3.8)
for pure water-methane system is shown in Figure 3.1 along with experimental
data (McLeod and Campbell, 1961; Marshall et al., 1964; Adisasmito et al., 1991;
Yang et al., 2001) and predictions from CSMHYD (Sloan and Koh, 2007). The
20
Pressure(MPa)
Seawater−methane (Model predictions)
15
70
Pressure (MPa)
10
60
5
50
40 0
270 275 280 285 290
Temperature (K)
30
20
10
0
270 275 280 285 290 295 300 305
Temperature (K)
Figure 3.1: Three-phase equilibrium curves for the methane-pure water system
compared with results from CSMHYD (dashed curve) and experimental data. Inset
shows experimental and calculated equilibrium curves for the methane-seawater
(0.6m NaCl solution) system.
CSMHYD program gives a good fit to experimental data at relatively low tem-
conditions. In comparison, our model gives much better agreement with experi-
The addition of dissolved ions shifts the equilibrium curve towards higher pres-
sures and lower temperatures (Figure 3.1, inset). The results are shown for 0.6m
NaCl solution, which has water activity approximately equal to that of seawater
(33.5 salinity). The results are in good agreement with the experimental values
21
(de Roo et al., 1983; Dickens and Quinby-Hunt, 1994) for three-phase equilibrium
cage occupancies of the small and large cages of structure I hydrate, which are
later used to calculate the solubility of methane in the aqueous phase in equilib-
rium with the hydrate phase (L-H zone). This helps to build an accurate and ther-
Thus, the calculation of solubility in the (L-H) region is most critical to the hydrate
distribution in the GHSZ. It is known from experimental results (Yang et al., 2001;
Seo et al., 2002; Servio and Englezos, 2002) and theoretical work (Handa, 1990;
Zatsepina and Buffett, 1998) that methane solubility in equilibrium with gas hydrate
Approximate models for computing methane solubility in the L-H region have been
reported (Servio and Englezos, 2002; Davie et al., 2004). However, we compute
proposed by Handa (1990). The final equations derived from the model are:
à ! P
∂ ln(xCH4 ) 1 Ve − V w − V CH4 2i=1 vi θi
= P2 xCH4 (3.9)
∂P T
RT i=1 vi θi − 1−xCH
4
à ! à ! à !
θi θi (xCH4 )P V CH4
ln = ln + ln + (P − Pe ) (3.10)
1 − θi P
1 − θi Pe
(xCH4 )Pe RT
where Ve is the molar volume of the empty hydrate lattice per mole of water, V w is
the partial molar volume of water, V CH4 is the partial molar volume of methane in
the aqueous phase and Pe is the three-phase equilibrium pressure at the tempera-
ture T . The values for Ve and V w are taken to be 22.5 cm3 /mol and 18.02 cm3 /mol
(Handa, 1990), respectively, while V CH4 is calculated using the relation reported
greater than the equilibrium pressure. However, they form a set of coupled, non-
summarized below:
using the numerical scheme outlined before. This calculation will also yield
• Use the equation of state (Duan et al., 1992) to compute the solubility of
methane (xCH4 )Pe in equilibrium with vapor at the three-phase point (Pe ,T ).
P.
23
As two examples, we show the relationship between methane solubility and pres-
data from Seo et al. (2002) is also shown for comparison. The peak in methane sol-
as pressure is increased.
0.0040
Methane solubility in aqueous phase (mole fraction)
0.0005
0.0
0 5 10 15 20 25 30
Pressure (MPa)
Figure 3.2: Variation of methane solubility in aqueous phase with pressure in the
L-V and L-H regions of the phase diagram at T = 278.15 K and T = 290 K, for
methane-pure water and methane-seawater (0.6m NaCl solution) systems.
24
otherwise stated, all results shown henceforth are for water with a constant water
0
Sealevel
5 0.001 0.495
10 0.0015 0.990
Pressure (MPa)
15 0.002 1.485
Depth (km)
20
0.0025
25 2.475
Seafloor
Hydrate + Water Geotherm
30 2.970
Base of GHSZ
Figure 3.3: Emergence of the GHSZ by imposing a seafloor depth, seafloor tem-
perature and geothermal gradient on the solubility contour plot (mole fraction). The
curve represents the three-phase seawater-methane equilibrium profile. A depth
scale corresponding to hydrostatic pressure (assuming constant seawater density
of 1030 kg/m3 ) is plotted on the right axis to apply this phase diagram to a marine
environment.
We now plot the contours of constant methane solubilities in the L-V and L-H
regions along with the three-phase equilibrium curve (Figure 3.3), using the mod-
25
els explained in the previous sections. In this example we assume the seafloor
is located at 2700 m below sea level, the seafloor temperature is 276 K, and the
geothermal gradient in the sedimentary section is about 0.016 K/m. This diagram
also shows the point where the temperature profile within the sediments intersects
the three-phase equilibrium curve at the depth marking the phase boundary, also
referred to as the base of the GHSZ. Thus, the region extending from the seafloor
down to the base of the GHSZ becomes the zone of hydrate stability (dissolved
methane and gas hydrate can coexist), whereas free gas can coexist with the aque-
ous phases below the GHSZ, provided methane in excess of the local solubility is
present.
Results from the contour plot can be combined into a single solubility curve
(Figure 3.4) that forms the foundation for accumulating gas hydrates in sediments.
In general, methane solubility increases with depth within the GHSZ, reaches a
local maximum at the base of the GHSZ and then slightly decreases below. The
specifics of the solubility curve, however, depend on water depth, seafloor temper-
The vertical depth below the seafloor (Figure 3.4) is normalized by depth to the
base of the GHSZ, denoted as Lt . The methane solubility is also scaled with the
at unit normalized depth with unit normalized methane solubility. The importance
of this normalization will be evident when we later show that this scaling makes our
1.5
Free gas + Water
2.0
2.5
3.0
0 0.5 1.0 1.5
Normalized methane concentration
Figure 3.4: Final methane solubility curve in a submarine setting obtained from the
phase equilibrium and methane solubility calculations. The depth scale is scaled
with respect to the depth of the base of the GHSZ and the methane concentration
is scaled with the peak solubility at the base of the GHSZ.
The phase stability and methane solubility curves are later used to predict hy-
up front, that this becomes a problem with our assumption of constant pore wa-
ter salinity during the simulations. Precipitation and dissociation of gas hydrate
changes the salinity, which, in turn, changes the phase boundaries and solubili-
Chapter 4
4.1 Introduction
that in situ (biogenic) generation of methane can contribute towards hydrate for-
mation on all continental margins and may be the principal source of methane
source of fluid flow. Freshly deposited sediments are unconsolidated, have high
porosity, low bulk density and high water content. However, with gradual burial un-
der younger sediments, the sediments lose their porosity, become denser and get
consolidated. The processes that bring about this change in sediment properties
are collectively grouped under the term diagenesis, which can be broadly divided
These diagenetic processes generally lead to the reduction of porosity and per-
meability, causing the bulk density to increase. This loss in porosity and sediment
fluid flow. The resulting porosity profiles are referred to as porosity-loss or com-
paction curves and are an indication of the sediment burial and subsidence histo-
In the absence of external fluxes, fluid flow in our system is assumed to originate
from mechanical diagenesis only. This is assumed because the flux of compaction-
driven fluid in a sedimentary basin is greatest in the upper few kilometers (Gal-
loway, 1984), which overlaps with the region where gas hydrates are normally
pressure conditions that are high enough to be excluded from our physical domain
of interest.
The theory of primary consolidation was first developed by Terzaghi (1943) by re-
lating the compaction and pore water expulsion to excess pore-pressures. Based
of a clay layer over time. Rubey and Hubbert (1959) explained the role of fluid pres-
29
and effective stress. The combined effect of vertical compaction-driven flow and
heat transfer was studied by Sharp (1976) for modeling rapidly subsiding basins.
Bethke (1985) for anisotropic and accreting domains by using a Lagrangian frame
sediments were reported to move upward toward the seafloor, whereas deeper
fluids tend to move laterally. A mathematical model for non-linear compaction was
proposed by Audet and Fowler (1992) in which different pore pressure regimes
tio of the Darcy flux to the sedimentation rate. Recent numerical models have
ments, pressure dissolution at grain contacts and multiple mineral reactions in this
hydro-mechanical problem (Luo et al., 1998; Yang and Fowler, 1998; Suetnova and
above this datum becomes a dependent variable (Bethke, 1985; Wangen, 1992;
Dugan and Flemings, 2000; Yang, 2001). As stated previously, we are interested
in modeling compaction-driven flow in the top most layer of the sediments. Hence,
a full basin-scale numerical model would not be a good choice for our problem,
problem, we fix our reference frame at the seafloor and assume that the seafloor
level stays constant over time. This assumption is commonly used in modeling
early (shallow) diagenesis (Berner, 1980) and helps us in writing the formulation
in Chapter 9.
fluid velocity profiles in the sediment. Porosity-loss can be empirically modeled us-
ing Athy’s law (Athy, 1930), which assumes an exponential decay of porosity with
sediment depth. However, it is well known that the driving force for porosity-loss
is effective stress and not sediment depth (Bear, 1988). A simple model for com-
rate
As explained before, the second assumption helps us to fix our reference frame at
the seafloor level, which stays constant in time. The depth co-ordinate is assumed
strain) and incompressible solid and fluid phases, the gradient of the total vertical
∂σv
= ρb g (4.1)
∂z
ρb = ρs (1 − φ) + ρf φ (4.2)
where ρs is the sediment density, ρf is the fluid density and φ is the porosity. Addi-
tionally we assume that the pore fluid pressure, pw , is hydrostatic so that the pore
and Bachmat, 1990) as the difference between the total vertical stress σv and the
32
pore pressure p:
σe = σv − p (4.4)
− σσe − σvσ−p
φ = φ∞ + (φo − φ∞ )e φ = φ∞ + (φo − φ∞ )e φ (4.5)
the freshly deposited sediments and φ∞ is the minimum porosity achieved at great
depths. The original relation has been modified to account for a non-zero minimum
à !
φ − φ∞ σe
ln =− (4.6)
φo − φ∞ σφ
Differentiating (4.4) with respect to z and using (4.1) and (4.3), we get:
∂σe
= (ρs − ρf )(1 − φ)g (4.7)
∂z
à ! à !
∂σe φo − φ∞ ∂ φ − φ∞
= −σφ (4.8)
∂z φ − φ∞ ∂z φo − φ∞
33
Eliminating the gradient of effective stress between (4.7) and (4.8) gives the follow-
à ! à !
φo − φ∞ ∂ φ − φ∞
−σφ = (ρs − ρf )(1 − φ)g (4.9)
φ − φ∞ ∂z φo − φ∞
4.3.1 Non-dimensionalization
It can be seen from equation (4.9) that the term σφ /[(1 − φ∞ )(ρs − ρf )g] has units of
σφ
Lφ = (4.10)
(1 − φ∞ )(ρs − ρf )g
z z
z̃ = = (4.11)
Lφ σφ /[(1 − φ∞ )(ρs − ρf )g]
The actual porosity along with the initial and final porosities are also rewritten as
φ − φ∞ φo − φ∞
φ̃ = , η= (4.12)
1 − φ∞ 1 − φ∞
In terms of these variables, (4.9) along with the boundary condition becomes:
1 ∂ φ̃
= −(1 − φ̃) , B.C. : At z̃ = 0, φ̃ = η (4.13)
φ̃ ∂z
34
Integration of the above equation leads to the following porosity profile along the
dimensionless depth:
η
φ̃ = (4.14)
η + (1 − η)ez̃
Sediment and fluid velocities can be easily obtained from the porosity-loss curves,
equations, it is generally assumed that the fluid and sediment velocities approach
Hutchison, 1985; Davie and Buffett, 2001). This is tantamount to saying that at
very low porosities, there is no relative motion between the sediment and pore-
fluid. The material balance for the sediment particles and fluid for this system can
be written as:
∂((1 − φ)ρs )
+ ∇ · (vs ρs (1 − φ)) = 0 (4.15)
∂t
∂(φρf )
+ ∇ · (vf ρf φ) = 0 (4.16)
∂t
where vs is the sediment velocity and vf is the fluid velocity. For time independent
densities and porosity profile φ(z), equations (4.15) and (4.16) become:
∇ · (vf ρf φ) = 0 (4.18)
35
which simply states that the flux of sediment and fluid each remains constant in
the system. The invariance of the sediment flux can be used to get the sediment
The assumption of solid and fluid velocities becoming equal at greater depths can
Utilizing relation (4.19) to get the sediment velocity vs,∞ , the fluid velocity is given
by:
Ṡ(1 − φo )φ∞
vf (z) = (4.21)
φ(1 − φ∞ )
Non-dimensionalizing the velocities with respect to the sedimentation rate and re-
duced porosities, the final expressions for the velocities are given by:
1−η
ṽs = (4.22)
1 − φ̃
1−η 1 − φ∞
ṽf = , where γ = (4.23)
1 + γ φ̃ φ∞
The above fluid balance is only performed for pore water in a compacting
medium, i.e. it assumes no gas hydrate or free gas formation and their effects
on the fluid flux. The complete water balance is presented in the next chapter.
36
Porosity and velocity profiles are computed for η = 6/9 and γ = 9, which corre-
Figure 4.1 shows how the porosity decreases monotonously from the maximum
the normalized depth z̃. Figure 4.2 shows sediment and fluid velocity profiles
1
Normalized depth
Parameter List
3 −−−−−−−−−−−−
η = 6/9
γ=9
4
6
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Reduced porosity
along the normalized depth. It can be seen that the sediment velocity is maximum
at the seafloor, where it is initially deposited, and then steadily decreases. The
fluid velocity, on the other hand, increases from its minimum at the seafloor and,
It is important to note that the fluid velocity is positive throughout the depth,
37
Normalized depth
2
Sediment velocity
Water velocity
3
Parameter List
−−−−−−−−−−−−
η = 6/9
4
γ=9
6
0 0.2 0.4 0.6 0.8 1
Normalized velocities
which means that the fluid always moves downwards relative to the seafloor. This
behavior is better understood by realizing that each new sedimentary layer being
deposited has generally more pore volume than the volume of water squeezed
out of the underlying layers, thus allowing some water into the system (de Caritat,
1989). Hence, although both water and sediment are carried into the subsurface on
the escalator of basin subsidence, the sediment in general moves faster, resulting
4.4 Conclusions
analytical expressions for sediment and fluid velocities were derived for the case of
hydrostatic pore pressures. The velocity profiles show that both fluid and sediment
always moves downwards with respect to the seafloor and upwards with respect to
the sediment. This fact is important in gas hydrate dynamics because it means that
Chapter 5
5.1 Introduction
Naturally occurring gas hydrates can derive their methane through two processes
with sediment can generate methane significantly faster than diffusion can remove
it toward the seafloor. This could lead to in situ precipitation of gas hydrate within
the GHSZ (Paull et al., 1994; Egeberg and Dickens, 1999; Davie and Buffett, 2001).
supply methane into the GHSZ from below (Xu and Ruppel, 1999; Haeckel et al.,
2004; Torres et al., 2004; Liu and Flemings, 2006). The progressive decrease of
methane solubility with shoaling depth across the GHSZ (Figure 3.4) implies that
gas hydrate can precipitate from methane charged fluids without requiring free
gas (Hyndman and Davis, 1992). The development of a general model needs to
consider gas hydrate accumulation from either of the two methane sources and
For in situ methanogenesis, solid organic carbon buried with sediments ulti-
proceeds via a series of complex reactions and pathways involving myriad com-
pounds (Berner, 1980; Claypool and Kvenvolden, 1983; Whiticar et al., 1986; Wall-
man et al., 2006). To simplify this process, we combine possible pathways for bio-
genic methane into a single first-order reaction (Martens and Berner, 1977; Davie
to the available organic carbon present in the sediments and linked to it through a
order model, we have neglected the complications arising from using such simple
kinetics for organic decay. More advanced models for simulating organic decay
and other coupled reactions have been recently published (Wallman et al., 2006),
but these introduce new parameters in the system, which makes our simulation
matter is actually the most efficient energy yielding metabolic process, but in the
absence of molecular oxygen, other oxidized substances can act as terminal elec-
tron acceptors to continue the degradation (Claypool and Kaplan, 1974). The
primary metabolic pathways, i.e. via reduction of carbon dioxide (respiratory pro-
et al., 1986). These methanogens can form methane from a limited number of
substrates, the most important of which are CO2 + H2 , acetate, formate, methanol
marine sediments, the evidence favors CO2 reduction as the dominant pathway
solved form as HCO3− ) reduction, with the following reaction pathway (Floodgate
CO2 + H2 O −→ HCO3− + H +
dioxide and hydrogen, followed by the reduction of CO2 to methane. When acetate
is the substrate, methane comes from the reduction of the methyl group, releasing
It was stated earlier that a variety of bacterial populations can metabolize the
organic substrate using different electron acceptors. When two or more types of
organisms compete for the same substrate supply, the one deriving the greatest
as their electron acceptor can compete so favorably with CO2 reducing bacteria
42
that methane production can be inhibited in the presence of dissolved sulfate ions.
dominant form of respiratory metabolism. The main zones that result from this
succession are:
Aerobic zone : Utilizes molecular oxygen as the electron acceptor, lying at the
Anaerobic sulfate reduction zone (SRZ) : Dissolved sulfate is used for oxidation
Anaerobic methane producing zone : Lies below the SRZ and contains the mi-
face (SMI), where the methane concentration goes to zero due to oxidation
by sulfate.
Due to this zonation, methane can only be present below the SRZ and conse-
mass fraction of total sediment. We note that represents only a fraction of the total
organic carbon (TOC) because not all can be converted to methane. The following
43
• sedimentation rate and the amount of degradable organic carbon at the seafloor
rate and organic carbon supply are not valid over geologic time-scales. This issue
above all gas hydrate systems examined to date (Borowski et al., 1999). The
SRZ reduces the amount of TOC available for methanogenesis. More crucially,
of the SRZ in our model shifts the zero methane boundary condition below the
seafloor, often by 5-20 meters, but only causes a small change in average hydrate
saturation. We therefore neglect the SRZ in the model and simulations presented
in this chapter. However, the utility of the depth of SRZ as a proxy for predicting
With the above assumptions, the material balance for degradable organic car-
∂ ∂
(ρs α(1 − φ)) + (ρs vs (1 − φ)α) = −ρs λα(1 − φ) (5.1)
∂t ∂z
44
The initial condition and boundary condition for this system are
The equations and variables are now rewritten in dimensionless form to reduce
also helps to distinguish gas hydrate systems into separate categories, such as
tem is better represented by the thickness of the GHSZ (Lt ), rather than the length-
z
z̃ = (5.4)
Lt
Further, the material balance can be written in terms of sediment and fluid
fluxes instead of velocities. The degradable content of organic matter in the sed-
iments is normalized with respect to its initial concentration at the seafloor (α0 ).
(Dm ). The sediment flux (Us ) is non-dimensionalized with respect to the fluid flux
α
α̃ = (5.5)
αo
t
t̃ = (5.6)
L2t /Dm
³ ´
1−φ∞
Us Ṡ(1 − φ0 ) φ∞
Uf,sed 1 − φ∞
Ũs = = = = =γ (5.7)
Uf,sed Uf,sed Uf,sed φ∞
λL2t
Da = (5.8)
Dm
Uf,sed Lt
Pe1 = (5.9)
Dm
lower methane diffusivity. The first Peclet number (Pe1 ) is the ratio of fluid advec-
tion to methane diffusion. The fluid flux used in defining Pe1 denotes the contribu-
Peclet number, defined later, that will quantify the effect of upward external flow
Using these dimensionless scalings, the organic mass balance (5.1) can be
rewritten as:
à !
∂ ³ ´ ∂ 1+γ
α̃(γ − φ̃) + Pe1 Ũs α̃ = −Da(1 − φ̃)α̃ (5.10)
∂ t̃ ∂ z̃ γ
The initial and boundary conditions for the dimensionless organic material balance
are:
Lφ , whereas the organic material balance equation (5.10) uses the GHSZ as a
characteristic length. To use the same expression for porosity in the organic bal-
ance equation, we have to define another dimensionless group (Ntφ ) which is the
Lt
Ntφ = (5.13)
Lφ
η
φ̃ = (5.14)
η + (1 − η)eNtφ z̃
47
We now develop equations for simulating gas hydrate and free gas accumulation
when dissolved methane concentration exceeds local solubility. In making this as-
of gas hydrate within marine sediments. Macroscopic specimens (>1 cm) of gas
hydrate have been recovered from drill cores in many locations (Kvenvolden, 1993;
Paull et al., 1996; Tréhu et al., 2003). However, analyses of pore water chemistry,
sediment temperature and downhole logs suggest that most gas hydrate in marine
diffusion dominates, but that it dissociates during core recovery (Paull et al., 1996;
Tréhu et al., 2003). We assume in our model that gas hydrate is disseminated
The following assumptions are made in formulating the methane material bal-
ance:
and in gas phase by cgm . The three-phase material balance can be written in one-
dimension as:
·
∂ h i ∂ φ
φ(1 − Sh − Sg )cwm ρf + φS h c h
m ρ h + φSg cg
m ρ g + Uf c w
m ρ f + Us Sh chm ρh
∂t ∂z 1−φ
¸ " #
φ g ∂ ∂cwm MCH4
+Us Sg cm ρ g = φ(1 − Sh − Sg )Dm ρf + ρs λαo (1 − φ)α̃ (5.15)
1−φ ∂z ∂z Morg
where ρf , ρh and ρg are densities of the water, gas hydrate and free gas, respec-
tively, Sh is the hydrate saturation in the pore space, Sg is the free gas saturation
in the pore space, and MCH4 and Morg are the molecular weights of methane and
organic matter, respectively. The first expression on the left side gives the rate of
accumulation of methane in the aqueous, gas hydrate and free gas phases. The
second term on the left side represents the flux of methane in the aqueous, gas
hydrate and free gas phases. The first term on the right side characterizes the dif-
fusion of methane in the pore fluid. The last term corresponds to the generation of
methane through the first-order organic reaction, and couples the methane mass
An initial condition and two boundary conditions are required to solve this par-
tial differential equation. We specify zero initial methane concentration in the pore
fluid, consistent with zero initial organic matter in sediment. A Dirichlet bound-
49
along with a Neumann boundary condition of zero diffusive flux of methane at the
bottom of the model (depth D). These conditions can be written as:
I.C. : cw
m (z, 0) = 0 (5.16)
B.C.(1) : cw
m (0, t) = 0 (5.17)
∂cw
m
B.C.(2) : (D, t) = 0 (5.18)
∂z
and scalings for depth, organic content and time as defined previously (equations
free gas phase (cgm ) are normalized with the methane solubility at the base of the
sults for a given seafloor depth, temperature, geothermal gradient and pore water
cw
m chm cgm
c̃w
m = , c̃hm = , c̃gm = (5.19)
cw
m,eqb cw
m,eqb cw
m,eqb
The dimensionless form of the three-phase methane balance can be written as:
" #
∂ 1 + γ φ̃ ³ ´
(1 − Sh − Sg )c̃w h g
m + Sh c̃m ρ̃h + Sg c̃m ρ̃g
∂ t̃ γ
·
1+γ ∂ 1 + γ φ̃
+ (Pe1 + |Pe2 |) Ũf c̃w
m + Pe1 Ũs
³ ´ (Sh
γ ∂ z̃ γ 1 − φ̃
50
¸ " #
∂ 1 + γ φ̃ ∂c̃w
h
c̃m ρ̃h + Sg c̃gm ρ̃g ) = Sw m
∂ z̃ γ ∂ z̃
à !
MCH4 ρs
+ βDa(1 − φ̃)α̃ (5.20)
Morg ρf
where Pe2 is defined in the next section, ρ̃h is the ratio of hydrate to fluid density,
ρ̃h is the ratio of free gas and fluid density, and β is the normalized initial organic
ρh ρg
ρ̃h = , ρ̃g = (5.21)
ρf ρf
αo
β= (5.22)
clm,eqb
The values of chm and ρ̃h are assumed to be constant and equal to 0.134 and 0.9,
respectively (Sloan and Koh, 2007; Davie and Buffett, 2001). Free gas density (ρ̃g )
is computed from the ideal gas law. Molecular weights MCH4 and Morg are taken
to be 16 and 30, respectively, while the ratio (ρs /ρf ) is set to 2.65/1.03 = 2.57. The
I.C. : c̃w
m (z̃, 0) = 0 (5.23)
B.C.(1) : c̃w
m (0, t̃) = 0 (5.24)
∂c̃wm
B.C.(2) : (D, t̃) = 0 (5.25)
∂ z̃
51
In gas hydrate settings where in situ methanogenesis provides all methane, the
fluid flux will be dominated by that resulting from compaction (Uf,sed ). However,
required. We thus superimpose an external flux (Uf,ext ) on the flux caused by com-
paction (Uf,sed ) to get a total flux (Uf ), as described by Davie and Buffett (2003b):
Since the depth variable is taken to be positive downwards, the flux due to sedi-
mentation remains positive. However, the external flux Uf,ext is directed upwards
and will have a negative value. If Uf,ext exceeds Uf,sed in magnitude, the net flux
will be negative, and fluid flow will be toward the seafloor. Alternatively, if the mag-
nitude of Uf,ext is less than Uf,sed , the net flux will be positive and fluid flow will be
down. In case of equal magnitudes, Uf will be zero and methane transport will
The total fluid flux is used as a boundary condition when solving the water mass
" #
∂ h i ∂ φ
φSw cw
w ρ f + φS h c h
w ρh + Uf c w
w ρf + Us Sh chw ρh = 0 (5.27)
∂t ∂z 1−φ
where cw h
w and cw are water mass fractions in the aqueous and hydrate phases,
respectively, and Sw is water saturation (Sw = 1−Sh −Sg ). The assumption of small
52
assumption of constant water density in the aqueous and hydrate phases makes
h i
the term φSw cw h
w ρf + φSh cw ρh invariant over time and drives the first term on the
left side of equation (5.27) to zero. This assumption helps to decouple the water
and methane mass balances, so that they do not have to be solved simultaneously.
" #
∂ φ φ
Uf + Us Sh chw ρ̃h = 0 =⇒ Uf + Us Sh chw ρ̃h = const (5.28)
∂z 1−φ 1−φ
As long as no hydrate forms in the system, the flux of water in the aqueous phase is
constant and equal to the total fluid flux. Once hydrate starts to accumulate in pore
space, water moves into the hydrate phase and the flux of water in the aqueous
à !
φ φ
Uf + Us Sh chw ρ̃h = Uf + Us Sh chw ρ̃h (5.29)
1−φ 1−φ z̃=0
dissolved methane (equation 5.23). This enables us to rewrite equation (5.29) as:
φ
Uf + Us Sh chw ρ̃h = (Uf )z̃=0 = Uf,sed + Uf,ext (5.30)
1−φ
53
φ
Uf = Uf,sed + Uf,ext − Us Sh chw ρ̃h (5.31)
1−φ
To develop the most general model applicable for all cases, such as zero sedimen-
tation, zero external flux or zero net flux (|Uf,sed | = |Uf,ext |), we normalize equation
(5.31) by the sum (Uf,sed + |Uf,ext |). Equation (5.31) can then be cast as:
Uf Uf,sed + Uf,ext Us φ
Ũf = = − Sh chw ρ̃h (5.32)
Uf,sed + |Uf,ext | Uf,sed + |Uf,ext | Uf,sed + |Uf,ext | 1 − φ
Multiplying numerators and denominators by the term (Lt /Dm ), and making use of
the definition of Peclet numbers, the fluid flux in dimensionless form is given by:
" #
Pe1 + Pe2 Pe1 Ũs (1 + γ φ̃)
Ũf = − Sh chw ρ̃h (5.33)
Pe1 + |Pe2 | Pe1 + |Pe2 | γ(1 − φ̃)
Uf,ext Lt
Pe2 = (5.34)
Dm
where Pe2 is a second Peclet number corresponding to the ratio of external flux
to diffusion. Thus, our numerical model has two independent Peclet numbers that
diffusion. Importantly, Pe1 will be positive, whereas Pe2 will be negative so that
fluid flux due to sedimentation dominates when |Pe2 | < |Pe1 | and external flux
dominates when |Pe2 | > |Pe1 |. For cases where external fluid flux dominates, the
tration, denoted by cw
m,ext , is imposed as a Dirichlet boundary condition at the bot-
tom of the domain, replacing the Neumann boundary condition in equation (5.25).
It is normalized as:
cw
c̃w
m,ext = m,ext (5.35)
cw
m,eqb
B.C.(2) : c̃w w
m (D, t̃) = c̃m,ext (5.36)
We first show the evolution of organic content within sediments (Figure 5.1) by
numerically solving the organic mass balance (equation 5.10). This hyperbolic
weighting was used for the advection term. All results shown henceforth are for
The profile of available organic matter in sediment depends on the ratio Pe1 /Da,
which given above (equations 5.8 and 5.9), signifies the ratio of sedimentation to
microbial decomposition during burial (Figure 5.1). Moreover, as the ratio Pe1 /Da
Base of
Normalized depth
1.0
GHSZ
2.0
3.0
0 0.2 0.4 0.6 0.8 1
Normalized organic concentration ~
α
a higher carbon input at the seafloor or a lower reaction rate. Solid organic carbon
available for methanogenesis reaches greater sediment depths with higher Pe1 /Da.
methane and water are solved numerically. The solution to the methane balance
ing the organic content profile in time. Solution of the equations in conservative
form ensures that there are no local material balance errors (all simulations had
56
which must be solved at each time step. This apparently under-determined system
can be solved, however, by realizing that all three cannot be the primary depen-
dent variable at any given time. As long as the methane concentration in pore fluid
and Sg are zero. Once dissolved methane exceeds solubility conditions within the
GHSZ, c̃w
m is set equal to the solubility, and excess methane goes into the hydrate
phase. In this case, Sg is set to zero, and Sh becomes the primary dependent vari-
free gas. In this case, Sh is set equal to zero, and Sg becomes the primary depen-
dent variable. Importantly, any gas hydrate crossing the base of GHSZ dissociates
to dissolved methane, which can become free gas provided pore fluid is already
primary dependent variable are done at each grid-block for each time-step of the
simulations.
ation with no external fluid flux. Transient profiles of dissolved methane concen-
57
tration and gas hydrate and free gas saturation can be obtained from numerical
simulations (Figure 5.2). Most simulations reach steady-state within 2-3 units of
dimensionless time (t̃). For Lt ≈ 450 mbsf (similar to Blake Ridge; (Paull et al.,
1996)) and Dm = 10−9 m2/s (Davie and Buffett, 2001), steady-state is achieved
within approximately 12 Myr. This residence time is similar to that predicted from
other mass balance calculations (Davie and Buffett, 2001; Dickens, 2003).
Simulations (Figure 5.2) follow a common path toward steady state. Initially,
are less than local solubility. When methane concentrations reach the solubility
curve, gas hydrate starts to precipitate. With additional methane supply and sedi-
ment burial, the fraction of gas hydrate in sediment increases, and the lowermost
occurrence of gas hydrate progressively moves down. In most cases, gas hydrate,
when formed, reaches the base of the GHSZ at steady-state. Gas hydrate dissoci-
ates upon crossing this horizon, generating free gas. Hence, at steady-state there
is a free gas layer immediately below gas hydrate (Davie and Buffett, 2001).
For both transient and steady-state simulations, a series of gas zones can exist
with dissolved methane and gas hydrate; and a lower zone with dissolved methane
and free gas. However, the presence and thickness of these gas zones can vary. A
more interesting case occurs with a modest reduction in the initial organic content
of the sediments (β). At relatively low β, depending on values for other parameters,
a steady-state can develop where gas hydrate does not extend to the base of the
GHSZ, and free gas does not form below the GHSZ (Figure 5.3). Essentially,
58
~ ~
Time t = 0.1 Time t = 0.1
0 0
Solubility curve
Hydrate saturation
Methane concentration
Normalized depth Free gas saturation
Normalized depth
0.5 0.5
Base of Base of
1.0 GHSZ 1.0 GHSZ
1.5 1.5
2.0 2.0
0 0.5 1.0 1.5 0 0.5 1
~ ~
Time t = 0.5 Time t = 0.5
0 0
Normalized depth
Normalized depth
0.5 0.5
Base of Base of
1.0 GHSZ 1.0 GHSZ
1.5 1.5
2.0 2.0
0 0.5 1.0 1.5 0 0.02 0.04
Steady state Steady state
0 0
Normalized depth
Normalized depth
0.5 0.5
Base of Base of
1.0 GHSZ 1.0 GHSZ
1.5 1.5
2.0 2.0
0 0.5 1.0 1.5 0 0.02 0.04 0.06
Normalized methane concentration Hydrate or free gas saturation
Figure 5.2: Simulations showing the time evolution of the dissolved methane con-
centration (left column) and the gas hydrate and free gas saturation in the sedi-
ments (right column). The results shown above are for methane generated from
biogenic in situ sources only. A finite gas hydrate layer along with a free gas layer
exists at steady state. The following model parameters were used for this simula-
tion: Pe1 = 0.1, Da = 10, β = 3, Pe2 = 0, η = 6/9, γ = 9. The seafloor parameters
used are for the Blake Ridge area: Seafloor depth = 2700 m, seafloor temperature
= 3◦ C and geothermal gradient = 0.04◦ C /m. We refer to these seafloor parameters
as the base-case values.
59
methane concentrations are less than the solubility curve starting at some depth
Xu and Ruppel (1999) have suggested that this case arises when methane sup-
ply from depth does not exceed a critical flux. Our simulations show that this situ-
ation can also occur when modest organic carbon input does not furnish enough
ical β has to be exceeded, analogous to the critical flux defined by Xu and Ruppel
Gas hydrate accumulation can also be simulated for systems where all methane
is supplied from deeper sources and β is zero (Figure 5.4). As seen from tran-
sient profiles, methane enters the system from below and forms gas hydrate within
the GHSZ when methane concentrations exceed solubility conditions. The gas hy-
hydrate extends to the base of the GHSZ and free gas forms below.
A certain minimum methane flux is required to form gas hydrate in this case.
However, a critical methane flux from below is not required for gas hydrate to con-
tact free gas at the base of the GHSZ. Assuming constant methane supply and
sufficient time, once hydrate has precipitated, a gas hydrate/free gas interface will
occur even at relatively low methane fluxes because sedimentation moves gas hy-
drate towards the base of the GHSZ. If we do not include sedimentation effects in
our model by setting Pe1 to zero, as done by Xu and Ruppel (1999), the hydrate
60
~ ~
Time t = 0.5 Time t = 0.5
0 0
Solubility curve Hydrate saturation
Methane concentration
Normalized depth
Normalized depth
0.5 0.5
Base of Base of
1.0 GHSZ 1.0 GHSZ
1.5 1.5
2.0 2.0
0 0.5 1.0 1.5 0 0.5 1
~ ~
Time t = 5.0 Time t = 5.0
0 0
Normalized depth
Normalized depth
0.5 0.5
Base of Base of
1.0 GHSZ 1.0 GHSZ
1.5 1.5
2.0 2.0
0 0.5 1.0 1.5 0 0.005 0.01 0.015
Steady state Steady state
0 0
Normalized depth
Normalized depth
0.5 0.5
Base of Base of
1.0 GHSZ 1.0 GHSZ
1.5 1.5
2.0 2.0
0 0.5 1.0 1.5 0 0.005 0.01 0.015
Normalized methane concentration Hydrate or free gas saturation
Figure 5.3: Simulations showing the time evolution of methane concentration and
gas hydrate saturation (similar to Figure 5.2), except that this system has a smaller
value of β, resulting in lesser carbon input. This causes an isolated hydrate layer to
exist at steady state, without any free gas below. The following model parameters
were used for this simulation: Pe1 = 0.1, Da = 10, β = 1.4, Pe2 = 0, η = 6/9, γ = 9.
Base-case seafloor parameters were used for this simulation.
61
~ ~
Time t = 0.1 Time t = 0.1
0 Solubility curve
0
Methane concentration
Normalized depth
Normalized depth
Base of Base of
1 GHSZ
1
GHSZ
2 2
Hydrate saturation
Free gas saturation
3 3
0 0.5 1.0 1.5 0 0.5 1
~ ~
Time t = 0.5 Time t = 0.5
0 0
Normalized depth
Normalized depth
Base of Base of
1 GHSZ 1 GHSZ
2 2
3 3
0 0.5 1.0 1.5 0 0.01 0.02 0.03
Steady state Steady state
0 0
Normalized depth
Normalized depth
Base of Base of
1 GHSZ 1 GHSZ
2 2
3 3
0 0.5 1.0 1.5 0 0.02 0.04
Normalized methane concentration Hydrate or free gas saturation
Figure 5.4: Simulations showing the time evolution of methane concentration and
gas hydrate and free gas saturation (similar to Figures 5.2 and 5.3), except that
the methane in this case is transported by an upward external flux from a deeper
source. A gas hydrate layer exists at steady state, with free gas below. The fol-
lowing model parameters were used for this simulation: Pe1 = 0.1, Da = 0, β = 0,
Pe2 = −2.0, η = 6/9, γ = 9. Base-case seafloor parameters were used for this
simulation.
62
phase becomes immobile within the sediments and we also observe the emer-
gence of a critical flux required to bring the hydrate and free gas layers together at
the base of the GHSZ (Figure 5.5). The methane solubility below the GHSZ was
assumed to be constant for comparison with the results of Xu and Ruppel (1999).
It can be seen that the critical flux needed for the hydrate and free gas phases to
0 0
Pe2 = − 1.0 Pe2 = − 2.0
Normalized depth
Normalized depth
0.5 0.5
1.0 Base of 1.0 Base of
GHSZ GHSZ
1.5 1.5
2.0 2.0
2.5 2.5
3.0 3.0
0 0.5 1.0 1.5 0 0.5 1.0 1.5
Normalized methane concentration Normalized methane concentration
0 0
Pe2 = − 4.0 Pe = − 5.0
Normalized depth
Normalized depth
0.5 0.5 2
Base of Base of
1.0 1.0
GHSZ GHSZ
1.5 1.5
2.0 2.0
2.5 2.5
3.0 3.0
0 0.5 1.0 1.5 0 0.5 1.0 1.5
Normalized methane concentration Normalized methane concentration
Figure 5.5: Steady state dissolved methane concentration profiles for the case of
methane supplied from deeper sources and without the effect of sedimentation
(Pe1 = 0). As Pe2 increases in magnitude (that is, upward external flux becomes
more important), the gas hydrate and free gas layers approach each other and
co-exist at the base of the GHSZ only if the critical flux Pe2 = −5 is exceeded.
Parameters used for these simulations are: Pe1 = 0, c̃w
m,ext = 1.0, Pe2 varies for the
four plots. Base-case seafloor parameters were used for this simulation.
63
Gas hydrate systems with in situ methanogenesis providing all methane can be
simulated by setting Pe2 = 0; those where rising fluids supply all methane can be
studied by setting β = 0. Gas hydrate systems with mixed sources of methane can
dent parameters. Thus, in the following two sections, we examine how changes in
various parameters affect the behavior of these two end member systems.
The simulations of gas hydrate systems considered so far (Figures 5.2, 5.3 and
5.4) pertain to particular sets of parameter values. To obtain a more general un-
Pe1 , Da and β dominantly affect the distribution of gas hydrate and free gas (Davie
of β and Pe1 can be divided into three fields (Figure 5.6). These fields define
conditions where gas hydrate will occur, and with or without free gas at the base of
the GHSZ. Paths can be followed across the fields to understand how parameters
affect gas hydrate systems. For example, consider a path of constant Pe1 (Figure
5.6): there is no gas hydrate at low β, gas hydrate without underlying free gas at
64
2.0
Parameter List:
1.8 −−−−−−−−−−−−
Da = 10.0
1.6 Pe = 0
2
Seafloor depth = 2.7 km
1.4 o
Seafloor temp. = 3 C
Geotherm = 0.04 oC/m
1.2
Pe1
0
0 0.2 0.4 0.6 0.8 1.0 1.2
β (Normalized organic content at seafloor)
Figure 5.6: Parameter space of Pe1 and β showing three different fields of gas
hydrate distribution. As β increases, with Pe1 held constant, the system moves
from the state of no gas hydrate to gas hydrate without free gas below and finally
to gas hydrate with free gas below.
intermediate β, and gas hydrate with free gas below the GHSZ at high β. The
separation of these three fields at relatively low Pe1 can be understood by realizing
that, at low Pe1 , the system is dominated by high rates of diffusion, which causes
loss of methane from the seafloor. To compensate for this loss, the burial of organic
gas hydrates.
As Pe1 increases, one might expect gas hydrates to form at progressively lower
values of β. However, this inverse relationship does not occur when Pe1 exceeds
65
about 0.6. Instead, the system requires more organic carbon (higher β) to form gas
hydrates (Figure 5.6). This effect results from the increase in sediment velocity
associated with the increase in Pe1 . Essentially, organic carbon moves through
the GHSZ faster, and because the Damkohler number is fixed, greater amounts of
2.0
1.8
1.6
1.4
1.2
Pe1
0.8
0.6
0.4
0.2
0 −1 0 1 2
10 10 10 10
β (Normalized organic content at seafloor)
Figure 5.7: Different gas hydrate regimes simulated for various ratios of Pe1 /Da.
The region to the left of each pair of curves represents no gas hydrate formation,
the narrow central region bounds the part where gas hydrate occurs without free
gas below, and to the right of each pair is the region where gas hydrates occur with
free gas below. Base-case seafloor parameters were used for this simulation and
Pe2 = 0.
that the ratio Pe1 /Da controls organic decay and subsequent methane generation.
66
Instead of keeping Da constant, therefore, the three distinct regions of gas hydrate
distribution can be examined for constant Pe1 /Da (Figure 5.7). Again, there are
region of hydrate without associated free gas at moderate β, and a region cor-
responding to hydrate with free gas below at high β. This representation serves
two purposes. First, curves representing gas hydrate formation and accumulation
steadily decrease in Pe1 as β gradually increases. The parameter space is, there-
fore, divided into two main regions: diffusion dominated at low Pe1 and advection
Ideally, single plots can summarize the dependence of gas hydrate systems
upon all relevant parameters. For in situ methanogenesis, this can be achieved
by realizing that each set of curves (Figure 5.7) represents different amounts of
organic carbon conversion; that is, the quantity that binds different curves together
is the net amount of methane generated within the GHSZ. At steady-state, this
quantity can be obtained from the normalized organic content at the base of the
GHSZ. The analytical solution to the organic mass balance equation (equation
5.10, at z̃ = 1) is:
h i −1
The net amount of methane generated within the GHSZ can be computed from this
67
If the curves shown in Figure 5.7 for different values of Pe1 /Da are rescaled and
plotted in Figure 5.8, with the x-axis now representing the methane generated
within the GHSZ (equation 5.38), we see that the different set of curves collapse
into one single pair of curves. The curves in Figure 5.7 covered ratios of Pe1 /Da
that were two orders of magnitude different. We are now able to summarize the
dependence of gas hydrate distribution over a wide range of the relevant parame-
ters (Pe1 ,Da,β) with just one plot. This clearly highlights the importance of scaling
A simple, generalized plot can also be constructed for systems where all methane
derives from deeper sources. The most important parameters in this case are the
methane sources come into effect only when |Pe2 | > |Pe1 |.
We first examine results for different values of Pe2 in the parameter space of
a zone of gas hydrate formation without free gas below at steady-state for this type
of source. Hence, there is only one curve for each value of Pe2 , which separates
parameters defining no gas hydrate formation from parameters defining gas hy-
68
2.0
1.8
1.6
1.0
Gas hydrate formation
0.8
without free gas below
0.6
0.4
No hydrate
0.2 formation
0
0 0.2 0.4 0.6 0.8 1
Organic carbon converted within GHSZ
Figure 5.8: Collapse of the curves shown in Figure 5.7 into one single pair of curves
achieved by rescaling and plotting them in a manner such that the x-axis represents
the total methane generated within the GHSZ (equation 5.38). Base-case seafloor
parameters were used for this simulation and Pe2 = 0.
drate with underlying free gas. Lower values of Pe1 represent diffusion-dominated
In the latter case, diffusive losses of methane are reduced, requiring less dissolved
methane in the external flux. Increasing the external fluid flux |Pe2 |, keeping Pe1
constant, implies greater net methane input to the system, thus requiring less c̃w
m,ext
To combine these multiple curves, we realize that net fluid flux is the controlling
factor. If we rescale the parameter space so that the y-axis represents the net fluid
69
−1
10
Hydrate formation
with free gas below
Pe2 = − 2
Pe1
0
10
Pe = − 4
2
Pe2 = − 6
No gas hydrate
formation
1
10
0.4 0.5 0.6 0.7 0.8 0.9 1
~
Cm,ext (Normalized methane concentration in external fluid)
Figure 5.9: Parameter space showing curves separating the region of no gas hy-
drate formation from hydrate formation with free gas below, for methane supplied
from deeper sources only (β = 0, Da = 0 and |Pe1 | < |Pe2 |). Base-case seafloor
parameters were used for this simulation.
flux in the system (algebraic sum of Pe1 and Pe2 ), the multiple curves for different
Pe2 collapse into a single curve (Figure 5.10). Thus, the entire parameter space for
methane supplied from deeper sources can also be defined in a single plot.
Different gas hydrate regimes can be delineated in two plots (Figure 5.8 and 5.10)
give any quantitative information about gas hydrate saturation (Sh ), the volume
70
−6
−5
Gas hydrate formation
Net Flux : Pe1 + Pe2
with free gas below
−4
−3
No gas hydrate
−2 formation
−1
0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
~
Cm,ext (Normalized methane concentration in external fluid)
Figure 5.10: Single curve separating region of no gas hydrate formation from gas
hydrate formation with free gas below by rescaling the y-axis of Figure 5.9, such
that it represents the net flux of fluid in the system. The curves corresponding
to different |Pe2 | in Figure 5.9 come together to yield a single curve for methane
supplied from deeper sources (β = 0, Da = 0 and |Pe1 | < |Pe2 |). Base-case
seafloor parameters were used for this simulation.
fraction of gas hydrate within sediment pore space. The parameter space contain-
ing gas hydrate in Figures 5.8 and 5.10 represent steady-state conditions. Thus,
each point inside these fields corresponds to a unique depth profile of gas hydrate
saturation. Most profiles (for example, Figures 5.2, 5.3 and 5.4) exhibit variable
gas hydrate saturation with depth. To get average gas hydrate saturation, we nu-
71
merically integrate the saturation profile over the depth of the GHSZ:
1 Z Lt
hSh i = Sh (z)dz (5.39)
Lt 0
2.0
1.8
1.6
1.4
1.2
Pe1
1% 2% 3% 4% 5% 6% 7%
1.0
0.8
0.6
0
0 1 2 3 4 5 6 7
Organic carbon converted within GHSZ
Figure 5.11: Average gas hydrate saturation (equation 5.38) contours plotted for
the case of biogenic methane generated in situ (Pe2 = 0). The set of curves plotted
in Figure 5.8 are also shown on the left side of the diagram. Base-case seafloor
parameters were used for this simulation.
Average gas hydrate saturations were evaluated for several simulations in the
gas hydrate forming region of Figures 5.8 and 5.10. The scaling schemes outlined
in previous sections assume even more importance when we observe that average
gas hydrate saturation also scales with the dimensionless groups and their com-
binations. We plot these saturation contours in Figure 5.11, for methane from in
72
situ sources, along with the set of curves defining the boundaries of the regions in
Figure 5.8. The average gas hydrate saturation within the GHSZ increases as Pe1
sized that this single plot of average hydrate saturation suffices for all values of the
Average gas hydrate saturations can also be evaluated for cases of external
methane supplied from deep sources. The average saturation contours, however,
do not scale if we apply the same scaling used to combine different curves in Figure
5.10. This happens because the y-axis in Figure 5.10 represents the total fluid flux
into the system, whereas hydrate saturation also depends on the rate at which
the sediments are moving. Higher sedimentation rate, characterized by large Pe1 ,
implies that the gas hydrate layer spends less time in the GHSZ, resulting in lower
dependent on Pe1 , which gets neglected if we take the y-axis to be the sum of
Pe1 and Pe2 . For example, if we simulate a case where Pe1 + Pe1 = −10, then
the average hydrate saturation will be different for the cases [Pe1 = 1,Pe2 = −11],
[Pe1 = 0.5,Pe2 = −10.5] and [Pe1 = 0.1,Pe2 = −10.1], although their sum remains
the same. The average gas hydrate saturations from the simulations for these
cases are 0.4%, 0.8% and 4%, respectively. Although the average gas hydrate
saturation is different for each case, we observe that the product of Pe1 and average
saturation remains constant. This means that the quantity that remains invariant is
the flux of the gas hydrate, which is simply the product of Pe1 and the average gas
hydrate saturation. Thus, the term Pe1 hSh i remains constant and scales with the
73
y-axis of Figure 5.10. The average saturation simulated for several parameters is
plotted as contours of constant Pe1 hSh i in Figure 5.12. The average saturation can
be calculated from these contours by dividing the contour value by Pe1 . It should be
noted that we always include the effects of sedimentation when modeling methane
from deeper sources. Thus, the results shown in Figure 5.12 hold for the case of
also observe a lower limit to the values of Pe1 for which the average gas hydrate
saturation scales with Pe1 . For very low values of Pe1 , the gas hydrate saturation
might approach 100% of the pore space, at which point the assumptions of our
model do not hold. The contours shown in Figure 5.12 are valid for Pe1 > 0.01
c̃w
m,ext in the external pore fluid. If the external methane concentration is slightly
more than the minimum required to precipitate gas hydrate, it takes much longer
All simulations discussed so far have been for constant seafloor depth, bottom
taken to be close to those for the crest of Blake Ridge (Paull et al., 1996). We refer
to these as the standard or base-case values, which are: seafloor depth of 2700
We now show that significant changes in these three parameters cause very
small changes to the average hydrate saturation contours shown in Figures 5.11
74
−50
−45 2.08 %
−40 1.84 %
Net Flux : Pe1 + Pe2
−35 1.59 %
−30 1.35 %
Cascadia Margin: Site 889
Fluid flow ~ 1 mm/yr
−25 Pe1 = 0.061
1.11 %
<S > = 3 %
−20 h 0.87 %
−15 0.63 %
−10 0.40 %
No hydrate formation
−5 0.17 %
0
~ 0 0.2 0.4 0.6 0.8 1
Cm,ext (Normalized methane concentration in external fluid)
Figure 5.12: Contours of gas hydrate flux Pe1 hSh i plotted along with the curves
separating the two regions of gas hydrate occurrence in Figure 5.10 for the case
of non-zero sedimentation and |Pe1 | < |Pe2 |. Gas hydrate saturation can be calcu-
lated by dividing the contour values with . For example, if Pe1 = 1, these contours
directly represent the average gas hydrate saturation in the pore space. Base-case
seafloor parameters were used for this simulation.
and 5.12 for different methane sources. The reason for this behavior is the manner
in which we non-dimensionalize the vertical depth in our model using the depth to
the GHSZ (Lt ) and methane concentration using the peak solubility at the base of
the GHSZ. Changes in seafloor parameters can cause a big change in the depth
of the GHSZ and peak methane solubility, thus causing a marked change in the
methane inputs required to form hydrates. But our scaling scheme transforms
these different methane solubility curves into curves that are very similar to each
75
Figure 5.13 (left column) shows the small change in average saturation con-
tours for biogenic in situ sources caused by large changes in seafloor temperature,
geothermal gradient and seafloor depth. The solid curves in each plot represent
the base case saturation contours while the dashed curves depict average sat-
uration contours for the perturbed system. It can be seen that average hydrate
Figure 5.13 (right column) also shows the similar effect of perturbations in
seafloor parameters on the saturation contours for methane supplied from deeper
sources only. For this case, the three subplots show the boundary separating the
region of hydrate with free gas below from the region of no hydrate for the base
case (solid) and the perturbed case (dashed). The change in the product Pe1 hSh i
from the base case is listed on the contours, where positive implies an increase in
value of Pe1 hSh i after applying the perturbation and negative indicates a decrease
in Pe1 hSh i. Thus, because of the small magnitude of these changes, our simulation
results shown in Figures 5.11 and 5.12 represent very good base cases, with the
parameters for any given geological setting acting as small perturbations to the
base case.
76
Pe1 + Pe2
1.5 perturbation: 3oC to 5oC
−30 − 0.10 %
1
3% 5% 7%
Pe
1.0 1%
−20 Seafloor temp. − 0.07 %
0.5 perturbation:
−10 o o − 0.04 %
3 C to 5 C
0 0
0 1 2 3 4 5 6 7 0 0.2 0.4 0.6 0.8 1
2.0 −50
Geotherm perturbation:
−40 − 0.08 %
Pe1 + Pe2
1.5 0.04 K/m to 0.08 K/m
−30 − 0.06 %
1
1% 7%
Pe
1.0 3% 5% Geotherm
−20 perturbation: − 0.04 %
0.5 0.04K/m to
−10 0.08K/m − 0.02 %
0 0
0 1 2 3 4 5 6 7 0 0.2 0.4 0.6 0.8 1
2.0 −50
Seafloor depth
−40 + 0.37 %
Pe1 + Pe2
1% 3% 5% 7%
Pe
Figure 5.13: Sensitivity of the average gas hydrate saturation contours to per-
turbations in seafloor temperature, geothermal gradient and seafloor depth. The
top plot in each column shows the temperature perturbation, the middle represents
geotherm perturbation and the bottom represents seafloor depth perturbation. The
left column shows contour plots depicting base case saturation contours (solid) and
those obtained after applying the perturbation (dotted) for biogenic in situ sources
only. The right column shows changes in Pe1 hSh i in response to these perturba-
tions for deeper methane sources. The solid curve represents the boundary be-
tween the region of hydrate with free gas and no hydrate for the base case, while
the dotted curve is the boundary after applying the perturbation. Change in values
of Pe1 hSh i compared to those shown in Figure 5.12 are also listed.
77
5.7 Implications
By sampling the entire parameter space for each type of methane source, our sim-
ulations reveal why natural gas hydrate systems can be so variable. Depending
on the methane inputs and outputs, there can be methane-charged sediment se-
quences containing no gas hydrate, gas hydrate but no free gas below the GHSZ,
or gas hydrate with free gas below the GHSZ. The saturations of gas hydrate and
free gas are also a function of the dimensionless parameters (Figures 5.11 and
5.12).
Some non-intuitive aspects of natural gas hydrate systems can also be under-
stood from our numerical modeling. For example, gas hydrates can precipitate
within the GHSZ from in situ sources of methane without a free gas layer below
even with continuous sedimentation. Previous modeling had simulated this sce-
nario, but only during the transient part of the simulations (Davie and Buffett, 2001).
produced is greater than the minimum needed to precipitate gas hydrates but less
than the amount required to extend the gas hydrate layer to the base of the GHSZ.
This finding has important implications because it suggests that gas hydrate sys-
tems can lack a gas hydrate/free gas contact, and hence a BSR on seismic profiles.
ODP Site 994 on outer Blake Ridge (Paull et al., 2000) may be an example of this
situation.
We have also shown that higher sedimentation rates do not necessarily imply
higher gas hydrate saturations, at least in the case of in situ methane sources.
78
With the initial TOC context fixed at the seafloor, and the reaction rate fixed with
time, increasing sedimentation rates can mean that a higher fraction of organic
carbon passes through the GHSZ. This could yield lower gas hydrate saturation at
steady-state.
Sedimentation rates and TOC input are not constant over geologic time-scales.
The saturation contour plots (Figures 5.11 and 5.12) provide a convenient way to
such changes. For example, if the sedimentation rate and TOC input over time are
known, one can plot points corresponding to different values of these parameters to
see how a particular gas hydrate system would evolve. This procedure, of course,
assumes that the gas hydrate system achieved steady-state after each change in
5.8 Conclusions
We have developed a numerical model for predicting gas hydrate formulation and
accumulation over geologic time-scales from methane generated either from bio-
genic in situ sources, methane from upward external fluxes or both. The following
in situ microbial activity can be divided into three categories: no gas hydrate,
gas hydrate without free gas below and gas hydrate with an underlying free
gas layer. The first Peclet number (Pe1 ), the Damkohler number (Da) and the
79
• For systems where methane derives from an external, deep source, the cat-
egory of gas hydrate without free gas below is not observed at steady-state
the time it takes to achieve steady state, but the steady-state system either
has no gas hydrate or gas hydrate directly underlain by a free gas layer at the
• Disconnected gas hydrate and free gas layers are observed in our exter-
nal flux simulations if we set sedimentation to zero (Pe1 = 0), because this
causes the gas hydrate and free gas layers to become immobile. Consistent
with the results of Xu and Ruppel (1999), a critical external flux has to be
exceeded to extend the gas hydrate and free gas layers to the base of the
GHSZ.
marized in two plots, one each for in situ and deeper methane sources, by
this happens when Pe1 is plotted against the amount of methane generated
within the GHSZ (equation 5.38). For deeper methane sources, this happens
when the net flux (Pe1 + Pe2 ) is plotted against the methane concentration of
• Average gas hydrate saturation contours also scale with dimensionless groups,
80
so that two contour maps provide gas hydrate saturation values for a large
ents, and are thus applicable to a wide variety of geological settings. Site-
2003a) become single points on these saturation maps. Our scaling schemes
make these plots ideal base cases for providing quantitative information about
the possible types of hydrate accumulation at any given location without per-
over most of the site-specific results available from hydrate modeling in the
literature, which are valid only for the numerous parameters relevant to a
Chapter 6
6.1 Introduction
The majority of natural gas hydrate accumulations are found in oceanic sediments
along continental margins where sufficient methane is available for gas hydrate
formation. Numerical models for gas hydrate accumulation in such settings have
been developed (Xu and Ruppel, 1999; Davie and Buffett, 2001; Gering, 2003;
Haeckel et al., 2004; Torres et al., 2004; Wallman et al., 2006), but their depen-
only to specific sites. Hence, most of these numerical models have been applied
to well studied sites, such as Blake Ridge or Hydrate Ridge (Cascadia Margin)
(Davie and Buffett, 2003a; Gering, 2003; Haeckel et al., 2004; Torres et al., 2004).
mal gradient, will result in different depths of the GHSZ and methane solubilities
within the GHSZ. This can lead to a different gas hydrate and free gas distribu-
tion at steady state, requiring numerical simulations for each parameter perturba-
into the physics governing these complex systems. For example, change in the
depth of the GHSZ by altering the seafloor depth, temperature, geothermal gradi-
82
ent or seawater salinity can change the gas hydrate system from being advection
to diffusion-dominated.
numerical model for simulating gas hydrate formation and accumulation over ge-
uration for a wide range of parameters into simple contour plots. One of these
the GHSZ, while the other plot represents the scenario where methane is supplied
from deeper sources via upward external fluxes (Bhatnagar et al., 2007).
We now compare our predictions with gas hydrate saturations inferred from
proxy data or other simulation studies for four well characterized gas hydrate set-
tings. We also show that incorporating mixed sources of methane can explain the
difference in gas hydrate and free gas distribution between Ocean Drilling Program
(ODP) Site 994 and Sites 995/997 at Blake Ridge. Finally, characteristics of sys-
tems with mixed sources of methane (both in situ biogenic and deeper sources), in-
cluding conditions suitable for relatively high gas hydrate saturations, are explained
The general features of the saturation contour plot for systems dominated by deep-
methane sources are discussed in the previous chapter. Here we show the utility of
83
this contour map by applying it to Site 889 (Cascadia Margin), which is part of a gas
and Davis, 1992; Westbrook et al., 1994; Davie and Buffett, 2003a).
TOC content (< 1%) and relatively high fluid flow rates (Westbrook et al., 1994;
Tréhu et al., 2003). Research cruises, including ODP Legs 146 and Leg 204, have
not only identified and quantified gas hydrate within sediments of this region, but
have also revealed some surprising aspects concerning gas hydrate distribution.
Layers of massive gas hydrate have been found at the southern summit of the Hy-
drate Ridge where methane actively vents from the seafloor (Tréhu et al., 2003;
Haeckel et al., 2004; Torres et al., 2004). These layers are not expected from most
numerical models for gas hydrate accumulation, which predict maximum gas hy-
drate saturations close to the base of the GHSZ (Xu and Ruppel, 1999; Davie and
Buffett, 2001). Focused fluid flow through coarse-grained conduits (Tréhu et al.,
2004; Milkov et al., 2005) and transport of methane in gas phase through the GHSZ
(Haeckel et al., 2004; Torres et al., 2004) has been hypothesized to cause such
issues regarding the presence and migration of free gas within the GHSZ. For ex-
ample, they may necessitate slow kinetic rates for gas hydrate precipitation and
gas dissolution in pore water (Haeckel et al., 2004; Torres et al., 2004) or salinity
Liu and Flemings, 2006, 2007). However, apart from specific sites near active
seafloor venting, sites in the area appear to be characterized by gas hydrate sat-
urations that increase with depth towards the BSR (Westbrook et al., 1994; Tréhu
et al., 2004).
We examine data from ODP Site 889 (Leg 146) west of Vancouver Island (West-
methane flux. According to our model, the average gas hydrate saturation for this
type of source can be constrained with the upward fluid velocity, sedimentation
rate, methane diffusivity, porosity profile, and other seafloor parameters listed in
Table 6.1. These dimensional variables can be combined into the dimensionless
groups, Pe1 and Pe1 + Pe2 , which are the main parameters controlling the gas hy-
Davie and Buffett (2003a) indicated that an upward velocity of about 0.42 mm/yr
(vf,0 , at the sediment interface) best fits the pore water chlorinity profile at Site 889.
This value is similar to that (∼1-2 mm/yr) believed to represent the region as a
whole (Wang et al., 1993). We use a velocity of 1 mm/yr at the seafloor to get the
net fluid flux (Uf = vf,0 φ0 , Table 6.1), which characterizes the sum of the two Peclet
numbers on the y-axis of the contour plot (Figure 6.1). The seafloor parameters at
this site lead to a predicted GHSZ extending to 233 mbsf, which compares favor-
ably with the BSR depth of 225 mbsf inferred from seismic data (Westbrook et al.,
1994). Calculations from the transport parameters give values for Pe1 + Pe2 and
Pe1 of -5.1 and 0.061, respectively (Table 6.1). This locates the point correspond-
ing to Site 889 at the contour close to Pe1 hSh i = 0.18% (Figure 6.1), and implies an
85
−50
−45 2.08 %
−40 1.84 %
Net Flux : Pe1 + Pe2
−35 1.59 %
−30 1.35 %
Cascadia Margin: Site 889
Fluid flow ~ 1 mm/yr
−25 Pe1 = 0.061
1.11 %
<S > = 3 %
−20 h 0.87 %
−15 0.63 %
−10 0.40 %
No hydrate formation
−5 0.17 %
0
~ 0 0.2 0.4 0.6 0.8 1
Cm,ext (Normalized methane concentration in external fluid)
Figure 6.1: Contours of gas hydrate flux Pe1 hSh i plotted along with the curves
separating the two regions of gas hydrate occurrence, for the case of non-zero
sedimentation and |Pe1 | < |Pe2 |. Gas hydrate saturation can be calculated by
dividing the contour values with . For example, if Pe1 = 1, these contours directly
represent the average gas hydrate saturation in the pore space
average gas hydrate saturation of 3%. The gas hydrate profile was also generated
(Figure 6.2), and shows gas hydrate saturation increasing toward 12% at the base
of GHSZ.
Numerical results from Davie and Buffett (2003a) for parameters close to those
listed in Table 6.1 indicate average hydrate saturation between 2 and 5%. As
explained in their discussion, these modest gas hydrate contents at Site 889 are
significantly lower than the 20-30% saturation above the BSR inferred from certain
analyses of seismic velocity and log resistivity (Yuan et al., 1996; Hyndman et al.,
86
Normalized depth
0.5 0.5
1.0 1.0
1.5 1.5
0 0.03 0.06 0.09 0.12 0 0.03 0.06 0.09 0.12
Hydrate saturation Hydrate saturation
Figure 6.2: Steady state hydrate saturation profiles, obtained using the parameters
listed in Table 6.1, at Blake Ridge, ODP Site 997 (left) and Cascadia Margin, ODP
Site 889 (right).
1999), but are consistent with the ∼ 3% value obtained from core temperature
We first describe general features of the average saturation contour plot obtained
for in situ biogenic methane sources. In such systems, self-similarity occurs when
we plot gas hydrate saturation contours as a function of Pe1 and the net amount
of organic carbon converted within the GHSZ (Figure 6.3). The net amount of
87
0.9
0.8
1
0.7
Peclet Number Pe
0.6
1% 2% 3% 4% 5% 6% 7%
0.5
0.4
Peru Margin
0.3 (b) Site 1230
<S > = 4.2%
Blake Ridge h
Peru Margin
0.2 Site 997 Site 1230
(a) Costa Rica <Sh> = 1.5% <Sh> = 4.9 %
Margin Blake Ridge
Site 997
0.1 Site 1040
<Sh> <1% <Sh> ~ 0 %
0
0 1 2 3 4 5 6 7
Organic carbon converted within GHSZ
Figure 6.3: Gas hydrate saturation contours averaged over the entire GHSZ for
systems where all methane is furnished through in situ biogenic reactions. Diffu-
sive losses dominate at low values of Peclet number (Pe1 ), implying that methane
generation within the GHSZ has to increase to form any gas hydrate. The set of
dashed curves, marked (a) and (b), represent the intermediate region of gas hy-
drate formation without free gas, with the region of no gas hydrate formation for
x-axis values lesser than for curve (a) and gas hydrate with free gas immediately
below for x-axis values greater than those for curve (b) (Bhatnagar et al., 2007).
Average gas hydrate saturation at different gas hydrate settings can be obtained
from this single contour map.
88
Table 6.1: Site-specific data, dimensionless groups and average gas hydrate satu-
ration calculation for Cascadia Margin and Blake Ridge sites.
Dimensionless groups
Uf,sed Lt
Pe1 = Dm 0.06 0.1 0.02
λL2t
Da = Dm — 2.1 2.1
U f Lt
Pe1 + Pe2 = Dm -5.1 0.1 0.02
(3/4)α0
β= clm,eqb
0 4.16 4.16
organic carbon converted within the GHSZ at steady state can be obtained from
the analytical solution of the organic balance equation and is written as (Bhatnagar
et al., 2007):
µ h i 1 ¶
1 − (1+γ)Pe1 /Da
= 1 − η + (1 − η) e β (6.1)
The set of dashed curves (Figure 6.3, (a) and (b)) separate the region of no gas
hydrate formation on the left side of (a) from the region of gas hydrate formation
with underlying free gas on the right side of (b). The narrow region enclosed by
this set of curves bounds the region where gas hydrate forms without any free gas
immediately below the GHSZ. This type of gas hydrate accumulation is discussed
in more detail later. Diffusive losses of methane dominate at lower values of Pe1 ,
requiring methane generation within the GHSZ to increase to form gas hydrate
(Figure 6.3). This can happen by either increasing the organic input to the sediment
(β) or allowing faster conversion of organic carbon into methane (lower Pe1 /Da)
(Bhatnagar et al., 2007). For the same amount of methane generated within the
GHSZ, increasing Pe1 causes the diffusive loss of methane to decrease, resulting
The utility of Figure 6.3 can be easily understood by realizing that this plot
summarizes information from numerous simulations performed over the entire pa-
rameter space. Thus, data from different geologic sites can be transformed into di-
mensionless variables defined above and mapped onto this contour map, yielding
90
average gas hydrate saturation at that site without any new numerical simulations.
Effects of perturbing the transport and seafloor parameters can also be directly
studied from Figure 6.3. We later apply this general contour plot to three different
Figure 6.3 shows that by gradually increasing β from a relatively low to high value,
for a given Pe1 , causes the system to go from no hydrate to an intermediate region
of hydrate with no free gas immediately below it and finally into the region of gas
hydrate with free gas contact at the base of the GHSZ. The narrow regime of gas
hydrate without free gas contact can be further divided into two subcategories. For
given Pe1 and Da, a relatively small value of β yields a steady state with an iso-
lated gas hydrate layer within the GHSZ without any free gas below (Figure 6.4a).
Increasing β causes an increase in the organic carbon input to the sediments and
causes free gas to form below the GHSZ, but not immediately below this boundary
(Figure 6.4b). Gas hydrate does not extend down to the base of the GHSZ in both
Simulation results in which both gas hydrate and free gas contact each other
at the base of the GHSZ (Davie and Buffett, 2001, 2003a; Bhatnagar et al., 2007)
show that free gas saturation equals some non-zero value just below the GHSZ.
This contrasts with the free gas saturation profile in Figure 6.4b, which starts to
increase from zero saturation at some depth below the GHSZ, because it does
not benefit from any excess methane input due to gas hydrate dissociation at the
91
0.5 0.5
Base of
1.0 GHSZ
1.0
1.5 1.5
2.0 2.0
2.5 2.5
3.0 3.0
0 0.5 1.0 1.5 0 0.005 0.010 0.015
(B) Pe1 = 0.1, Da = 10, β = 1.42
0 CH solubility curve
0
4 Gas hydrate
CH4 concentration Free gas
Normalized depth
0.5 0.5
Base of
1.0 GHSZ 1.0
1.5 1.5
2.0 2.0
2.5 2.5
3.0 3.0
0 0.5 1.0 1.5 0 0.005 0.010 0.015
Normalized methane concentration Gas hydrate/free gas saturation
Figure 6.4: Different cases of gas hydrate formation without free gas immediately
below the hydrate layer. (A) For given Pe1 and Da, a relatively lower value of β im-
plies lesser organic carbon input to the system, resulting in gas hydrate formation
without any free gas below it. (B) Increasing β by a small amount raises organic
carbon input such that free gas forms beneath the hydrate layer but does not con-
tact it at the base of the GHSZ. The gas hydrate layer itself does not extend to the
base of the GHSZ in both the cases. Further increase in β would bring the gas
hydrate and free gas layers together at the base of the GHSZ.
92
base of the GHSZ. Thus, occurrence of a BSR for such cases is unlikely due to two
reasons. First, there is no gas hydrate/free gas contact. Secondly, increasing β can
bring together hydrate and free gas close to the base of the GHSZ, but because
free gas saturation gradually increases from zero, the resulting impedance contrast
In general, average gas hydrate saturations for this special category of no free
gas immediately beneath the GHSZ are low for Pe1 >∼ 0.05 (Figure 6.3). This can
be inferred from the distance between the 1% contour and the dashed curve (a)
in Figure 6.3, which represents the 0% contour as it marks the boundary where
gas hydrate first forms. However, for Pe1 <∼ 0.05, the contours start to converge
together such that average saturation within this narrow regime starts increasing
to significant values. Hence, appreciable amounts of gas hydrate can form without
free gas immediately below the GHSZ only for small values of Pe1 , which physically
translates into geologic settings with low sedimentation rates. This relationship is
used to model gas hydrate distribution at one of the Blake Ridge sites.
The Blake Ridge system is probably the most extensively studied marine gas hy-
drate province, with ODP Leg 164 providing detailed information about sediment
properties, pore-water and gas composition, gas hydrate samples, well-log data
and seismic profiles (Paull et al., 2000). This region is a passive margin setting
characterized by low advection rates (Egeberg and Dickens, 1999) and modest
organic content, with total organic carbon (TOC) between 0.5 - 1.5 wt.% (Site
93
997) (Paull et al., 2000). Linear sulfate gradients within the sulfate reduction zone
(SRZ) of the Blake Ridge sediments indicate that the system is diffusion dominated
(Borowski et al., 1996). Pore water Br− and I− profiles also suggest low advection
rates (Egeberg and Dickens, 1999). Geochemical analysis of gas and other pore
water constituents imply that the methane is of biogenic origin. These results in-
dicate that the Blake Ridge system is dominated by in situ generated methane,
which is also supported by the success of matching the chloride anomaly at Site
997 with that predicted from numerical models using in situ methane sources only
be the only source, with flow rates obtained from pore-water bromide-iodide ratios
(Egeberg and Dickens, 1999), no gas hydrate formation occurs at Site 997 (Davie
and Buffett, 2003a). Hence, we model Site 997 as a gas hydrate setting where in
Simulating gas hydrate accumulation due to this methane source requires spec-
ification of the Peclet number, Pe1 , the dimensionless organic carbon content at the
seafloor, β, the Damkohler number, Da, and the seafloor parameters, i.e. poros-
ity, seafloor depth, temperature and geothermal gradient. The definitions of these
Table 6.1 for Blake Ridge Site 997. The TOC at Site 997 is assumed to be 1.5
wt.% (Paull et al., 2000; Davie and Buffett, 2001). The fraction of labile TOC that
is actually converted to methane is usually lesser than unity and a value of 50%
94
simulation results of Davie and Buffett (2001), we assume available TOC to be 75%
ration contours in the form shown in Figure 6.3 is the ability to track the dynamic
evolution of gas hydrate systems over geologic time scales. This evolution can be
and/or TOC content at the time of burial. Changes in seafloor parameters such
as seafloor depth, bottom water temperature and geothermal gradient can also
be incorporated through their effect on the depth of the GHSZ (Lt ), which is an
intrinsic parameter in the dimensionless groups. Plotting these points on the con-
tour map and tracking their loci can then give information about how any given gas
hydrate province has evolved in time. To illustrate this point, we plot two points
for the lower Pliocene and Miocene sequences and 4 cm/kyr for the entire Qua-
ternary sediments at Site 997 (Paull et al., 1996; Davie and Buffett, 2001). The
point corresponding to the earlier high sedimentation rate indicates average gas
hydrate saturation of about 1.5%, while the relatively recent sedimentation rate
moves the point out of the saturation contour region and into the regime of no gas
hydrate formation (Figure 6.3). TOC contents for both sequences were assumed
to be the same. Thus, the gas hydrate system at Blake Ridge is probably evolv-
ing from higher saturation towards a steady-state that is devoid of any gas hydrate.
Moderate sedimentation rates in the past, along with modest TOC content, allowed
95
greater amounts of gas hydrate to accumulate, but this level of saturation cannot
be sustained by the current influx of organic carbon. Hence, this gas hydrate sys-
tem might be gradually losing more methane to diffusive losses than is being fed
to it through methanogenesis.
Average gas hydrate saturation predicted from our contour plot using sedimen-
tation rate of 22 cm/kyr (∼1.5% of pore space) matches well with an average of
2% inferred from numerical simulations of Davie and Buffett (2001) for the same
parameter set. It should be noted that the saturation predicted from the contour
plot (Figure 6.3) is averaged over the entire GHSZ, while values presented in the
literature usually report average gas hydrate saturation over the hydrate occur-
rence zone only. The gas hydrate saturation profile simulated for these parameters
is shown in Figure 6.2 and indicates peak saturation of about 6% at the base of
the GHSZ. This matches well with the results of Davie and Buffett (2001) that pre-
dict peak hydrate saturation of about 5%. Gas hydrate saturation averaged over
the hydrate occurrence zone from our simulation (Figure 6.2) yields a saturation
uration within the the hydrate occurrence zone to be in the range 2-6% (Holbrook
et al., 1996; Egeberg and Barth, 1999; Collett and Ladd, 2000; Lee, 2000). Hence,
our simple one-dimensional model and the associated contour plot is able to give
Heterogeneous gas hydrate/free gas distribution at Sites 994, 995 and 997
The point corresponding to Site 997, shown in Figure 6.3, lies in the region charac-
terized by free gas immediately below the base of the GHSZ. However, this might
not be representative of the entire Blake Ridge region. The actual distribution of
gas hydrates and free gas at various sites might be quite different from each other
qualitatively as well as quantitatively. This is evident from the patchy BSR and free
gas distribution at Sites 994, 995 and 997. While gas hydrate and free gas extend
as connected phases up to the base of the GHSZ at Site 997, free gas appears
to be disconnected below the GHSZ at Site 995 and does not exist directly below
the GHSZ at Site 994 (Holbrook et al., 1996; Paull et al., 2000). Relatively low fluid
flux, compared to that at Sites 995 and 997, has been hypothesized to prevent
gas hydrate and free gas from extending to the base of GHSZ at Site 994 (Xu and
Ruppel, 1999). The model of Xu and Ruppel (1999), however, assumes no sed-
from deeper sources only. On the other hand, modest sedimentation rates and
TOC contents across this region suggest that biogenic input of methane cannot be
sedimentation) through which the absence of a BSR at Site 994 and occurrence of
For all sites it is assumed that the biogenic input of methane, and the parame-
ters that control this source, remain invariant. This can be justified by the similar
sedimentation rates, TOC contents and seafloor parameters (i.e., seafloor depth,
97
temperature and geotherm). For simplicity, we also assume that the depth to the
base of the GHSZ, Lt , is the same for all three sites. Hence, the dimensionless
We commented previously that gas hydrate without free gas immediately be-
low can accumulate in substantial amounts through biogenic in situ sources at low
Peclet numbers (Pe1 ) only. Thus, if gas hydrate forms only through this methane
input without free gas contact at Site 994, a relatively small Pe1 will be required.
The sedimentation rate used for Site 997 (Figure 6.3) varies between 22 cm/kyr
and 4 cm/kyr, with the average rate between these limits. We choose an aver-
age sedimentation rate of 11 cm/kyr, which is in agreement with the value of 12.2
cm/kyr used by Egeberg and Dickens (1999) to model gas hydrate distribution at
Site 997. Using this rate results in Pe1 equal to 0.05. In calculating the normalized
organic carbon input in Table 6.1, we assumed that only 75% of this TOC is avail-
able for methanogenesis, which was done to compare simulation results with those
reported by Davie and Buffett (2001). This value of the actual fraction of available
organic carbon is poorly constrained and, in general, lower than 75%, with values
in the range 25-50% being more common (Henrichs and Reeburgh, 1987; Davie
and Buffett, 2003a; Buffett and Archer, 2004). Consequently, we assume an in-
termediate value of 40%, which combined with a 1.5 wt.% TOC content yields a
normalized organic carbon input of β = 2.2. The Damkohler number, Da, remains
This biogenic input remains constant for all three sites, but it is known that Sites
995 and 997 have free gas immediately below gas hydrate at the base of the GHSZ
98
(Holbrook et al., 1996). Hence, the only other methane source that can be invoked
to bring the hydrate and free gas layers together at the base is an upward external
fluid flux. This is also justified by previous modeling studies that show Site 997
to be characterized by low upward flux (Egeberg and Dickens, 1999; Hesse et al.,
2000) and that the chloride anomaly at Site 997 can be matched well with a mixed
methane source (Davie and Buffett, 2003a). Egeberg and Dickens (1999) used
an upward flow rate of 0.2 mm/yr to fit pore water halide concentrations, thereby
constraining the gas hydrate saturation at Site 997. Davie and Buffett (2003a)
net upward velocity of 0.08 mm/yr at the seafloor. An external upward fluid flux is
easily introduced in our model by specifying a negative value for Pe2 . The resulting
net fluid flux through the system can be characterized by adding the two Peclet
numbers, i.e., the net Peclet number for this mixed source system is Pe1 + Pe2 . We
use this seafloor velocity (0.08 mm/yr) as the net upward external flux through the
system, which transforms into Pe1 + Pe2 = −0.86, resulting in Pe2 = −0.81.
Simulations are now performed for these two sets of parameters: one with bio-
genic in situ methane generation only (Figure 6.5a) and the other with a mixed
thesis, Figure 6.5 shows a dimensional depth axis for comparison with field values.
Figure 6.5a shows that at steady state and in the absence of any upward flux at
Site 994, gas hydrate remains isolated within the GHSZ without any free gas con-
tact directly below the hydrate. Free gas does form in the system starting at about
550 mbsf, which is in agreement with the depth inferred from vertical seismic pro-
99
Pe = 0.05 Pe = 0.05
1 1
100 100
Da = 2.1 Da = 2.1
β = 2.25 β = 2.25
200 Pe = 0 200 Pe = − 0.86
2 2
300 300
Depth (mbsf)
Depth (mbsf)
400 400
Base of Base of
GHSZ GHSZ
500 500
600 600
700 700
800 800
0 0.01 0.02 0.03 0 0.02 0.04 0.06 0.08
Gas hydrate/free gas saturation Gas hydrate/free gas saturation
Figure 6.5: Steady state heterogeneous gas hydrate (solid curves) and free gas
(thick dashed curves) distribution at Sites 994, 995 and 997 (Blake Ridge). The
biogenic parameters remain constant for both cases. Pe1 = 0.05 and β = 2.25
listed in the insets correspond to an average sedimentation rate of 11 cm/kyr and
TOC content of 1.5 % (with 40 % available organic carbon), respectively. Other
seafloor parameters and constants are the same as those listed in Table 6.1 for
Blake Ridge Site 997. (A) Site 994, simulated only with a biogenic in situ methane
input, shows an isolated gas hydrate layer that does not extend to the base of the
GHSZ. Free gas starts at 550 mbsf, leaving a 90 m zone beneath the GHSZ devoid
of any free gas, thus explaining the absence of any BSR at this site. (B) Using the
same biogenic methane input as in case (A), but with a low upward flux recycling
methane back into the GHSZ, gas hydrate as well as free gas extend to the base
of the GHSZ at Sites 995 and 997, resulting in BSRs at both these locations.
100
filing (VSP) data (Holbrook et al., 1996). Gas hydrate saturation averaged over the
GHSZ at Site 994 is 0.5%, averaged over the hydrate occurrence zone is 1.2% and
the peak saturation close to the base of the GHSZ is 2.1%. For Sites 995 and 997,
we use identical upward fluid flux in addition to in situ methane generation, which
gives the same saturation profiles for both sites. Figure 6.5b indicates a steady-
state where gas hydrate is in contact with free gas at the base of the GHSZ, with
average saturation over the GHSZ, over the hydrate occurrence zone and peak
Site 995 has smaller amounts of gas hydrate compared to Site 997 (Paull et al.,
2000), which can be simulated by lowering the net upward fluid flux at this site. This
would lower the average gas hydrate saturation, but free gas still forms immediately
beneath the hydrate layer. However, due to lack of fluid flow rate data at Site 995,
we approximate it by the same flux as that present at Site 997. Free gas beneath
the GHSZ occurs as a disconnected phase at Site 995, which cannot be simulated
using our simple one-dimensional model. Such features are probably governed
modeling in at least two spatial dimensions. Although our model cannot resolve
good average estimates of average gas hydrate saturation at these three sites in
The Peruvian continental margin has been the subject of two previous ODP Legs.
While the primary objective of ODP Leg 112 was to study the tectonic evolution
of the margin, the more recent Leg 201 was dedicated towards studying microbial
life deep beneath the seafloor. Gas hydrates were recovered at two sites cored
during Leg 112 (Site 685 and 688), in addition to geochemical evidence for hy-
drates and BSR occurrence at most of the drilled sites (Kvenvolden and Kastner,
1990). Chloride excursions near the base of the holes indicate hydrate saturation
of 10%, 9%, 3% and 9% at Sites 682, 683, 685, and 688, respectively (Kvenvolden
and Kastner, 1990). Molecular composition of hydrocarbon gases and the car-
bon isotopic compositions of methane and carbon dioxide indicate that microbial
mediated processes are primarily responsible for methane in the hydrates (Suess
and von Heune, 1988; Kvenvolden and Kastner, 1990). An interesting feature of
the Peruvian continental margin sediments is the consistently high organic carbon
nary section (Suess and von Heune, 1988). For example, TOC content at Site
679 ranges between 1.5-7.8% (averaging ∼ 4%) in the Quaternary section and
between 1.5-6.5% (averaging ∼ 4.7%) in the early Pliocene interval (ten Haven
et al., 1990). These observations, along with the high sedimentation rates, clearly
indicate a high carbon influx into the system. Additionally, this large carbon input
appears to be relatively recent, which implies that this hydrate system might not
Table 6.2: Site-specific data, dimensionless groups and average gas hydrate satu-
ration calculation for Costa Rica Margin and Peru Margin sites.
Dimensionless groups
Uf,sed Lt
Pe1 = Dm 0.04 0.19 0.08
λL2t
Da = Dm 1.4 5.1 5.1
(3/4)α0
β= clm,eqb
2.5 6.8 6.8
During Leg 201, Site 1230 was drilled through hydrate bearing sediments on
the lower slope of the Peru Trench to study how microbial activities in hydrate se-
quences differ from sediments that lack hydrate. This site lies close to Site 685 of
Leg 112 and is also characterized by modest sedimentation rates (10-25 cm/k.y.)
and TOC rich sediments (2-4%) (D’Hondt et al., 2003). Pore water analyses at
Site 1230 indicate high microbial activity through elevated DIC, alkalinity and am-
two different sedimentation rates to model Site 1230 and assume an average TOC
content of 3 wt.%, with 75% of this organic carbon being available for methanogen-
We explained the use of our contour plot in tracing the geologic history of any
gas hydrate province in the previous section (Blake Ridge). As another example
of this feature, we plot two points (Figure 6.3) corresponding to different average
sedimentation rates of 25 cm/k.y. for the Miocene sequence and 10 cm/k.y. for the
Pleistocene sequence at Site 1230 (D’Hondt et al., 2003). The corresponding Pe1
are 0.19 and 0.08, respectively. With the same TOC for both sedimentation rates,
the net amount of methane generated within the GHSZ is 4.2 and 4.9 normalized
units, respectively (Table 6.2). On plotting these points on the contour plot (Fig-
ure 6.3) we note that the point corresponding to the relatively recent Pleistocene
sedimentation rate indicates average gas hydrate saturation of about 4.9%, while
the Miocene sedimentation rate yields an average saturation close to 4.2%. Thus,
104
the gas hydrate system at Peru Margin is probably evolving from a lower average
This feature is non-intuitive in the sense that lower sedimentation rate would
generally imply a lower organic carbon influx to the system, but in this case yields
by realizing that decreasing sedimentation rate also decreases the sediment ve-
locity through the GHSZ, thereby allowing more time for the organic carbon to fully
tion rate beyond some critical value leads to higher average gas hydrate saturation
The analysis presented above assumes that the gas hydrate system has achieved
steady state after applying perturbations to the sedimentation rate. However, the
high TOC input to the sediments in the Peru Margin appears to be relatively recent,
implying that this system might not have achieved steady state. Hence, the aver-
age gas hydrate saturation predicted at the recent sedimentation rates might be
higher than that inferred from log data at Site 1230. Interstitial water chloride ex-
cursions at 80 mbsf and 150 mbsf at Site 1230 coincide with the two intervals from
which solid hydrate samples were recovered. These chloride anomalies suggest
gas hydrate saturation of about 20% in the pore space (D’Hondt et al., 2003).
ODP Leg 170 was dedicated towards studying heat, mass and fluid fluxes in
the accretionary complex formed by the subduction of the Cocos plate beneath
105
the Caribbean plate. Upper plate sediments drilled at Site 1040 indicate a shal-
low sulfate depleted zone, methane saturated pore waters beneath this SRZ and
et al., 1997). In contrast, the underthrust sediments are rich in sulfate, have
low methane content and exhibit a second sulfate reduction zone beneath the
décollement (Kimura et al., 1997). TOC contents in the upper plate sediments
at Site 1040 vary slightly, with a mean value of about 1 wt.% (Kimura et al., 1997;
Hensen and Wallman, 2005). A large ammonia peak observed at about 150 mbsf
indicates high rates of organic decay and in situ methane formation. These ob-
servations reflect the importance of methane generation within the upper plate
sediments from organic carbon decay and consequently we model the Costa Rica
to the given geologic setting, lateral and cross-décollement fluid flow can occur due
possible effects of such fluid fluxes along with in situ methane generation on the
amount of gas hydrate. Application of contour plot (Figure 6.3) to Site 1040 can
not include these mixed sources, but will help constrain the amount of gas hydrates
averaging 7.2 ◦ C/km, which leads to the GHSZ extending over the entire sedi-
ment column (Kimura et al., 1997). This leads to the absence of any free gas
zone beneath the GHSZ within considerable sediment depth, which might explain
the absence of any BSRs on the seismic sections obtained in the Leg 170 study
106
area. Our model, however, requires specification of the GHSZ depth, Lt , to cal-
equal to the décollement depth (∼ 370 mbsf), which can be justified by low pore
water methane contents and absence of other gas hydrate proxies beneath the
décollement (Kimura et al., 1997). Thus, we assume Lt = 370 mbsf for the cal-
2005). This leads to Pe1 = 0.04, Da = 1.4 and normalized TOC input of β = 2.5
(Table 6.2). In this case the amount of available organic carbon within the TOC is
assumed to be 25% for comparison with the results of Hensen and Wallman (2005)
for a similar parameter set. On calculating the net amount of methane generated
within the GHSZ, the point corresponding to the Site 1040 parameters is plotted in
Hensen and Wallman (2005) simulated various scenarios to predict gas hydrate
saturation at Site 1040, one of them being the case where all methane derives
from in situ organic decay. This scenario yields average gas hydrate saturation
of 0.8% over the hydrate occurrence zone using transport parameters similar to
those stated above. Gas hydrate saturation averaged over the entire GHSZ is al-
ways less than that averaged over the occurrence zone, and can be a factor of
two or three smaller than the latter value. This indicates an average value over the
occurrence zone of about 0.4-0.6% from our method, which is in agreement with
the value predicted by Hensen and Wallman (2005). It is also worth mentioning
that Site 1040 in Figure 6.3 lies within the region of gas hydrate formation without
free gas contact at the base of the GHSZ, which indicates that the gas hydrate
107
layer does not extend down to the décollement. Interestingly, some of the scenar-
ios simulated by Hensen and Wallman (2005) that included upward fluid flow also
indicated this feature. However, absence of free gas beneath the décollement at
Site 1040, as predicted by its location in the contour plot, should not be understood
as a direct result of our model. In other words, absence of free gas is due to the
GHSZ extending all the way to the basement and not due to insufficient supply of
methane.
The contour plot (Figure 6.3) summarizes qualitative and quantitative informa-
tion about gas hydrate saturations in marine sediments under a host of different
general predictions about conditions where gas hydrate might accumulate in large
amounts. We first discuss such conditions for hydrate settings with only deeper
methane sources (due to their simplicity) and then generalize them for systems
For gas hydrate formation from deeper sources, the analysis is relatively straight-
forward. Higher net fluid fluxes (Pe1 + Pe2 ) will give higher gas hydrate saturations,
while higher sedimentation rates characterized by Pe1 will result in lower average
saturation. Gas hydrate saturation is thus proportional to the ratio of these two
108
groups, that is (Pe1 + Pe2 )/Pe1 . Taking this ratio cancels the GHSZ thickness Lt
and methane diffusivity from the expressions, so that average saturation becomes
a function of the external flux and sedimentation rate only. The thickness of the
gas hydrate layer within the GHSZ, however, is only a function of the net flux.
0.2 Pe = 0.10
1
Pe1 = 0.25
0.4 Pe1 = 0.50
Normalized depth
0.6
0.8
1.0
1.2
1.4
1.8
2.0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
Hydrate saturation
Figure 6.6: Effect of burial rate, as characterized through the Peclet number Pe1 ,
on steady state gas hydrate saturation profiles from deeper sources of methane.
The three simulations represent different Pe1 for the same net fluid flux Pe1 + Pe2 .
Increasing Pe1 results in higher sediment velocities, shorter residence times within
the GHSZ and lower gas hydrate saturation. Depth to the top of the gas hydrate
layer depends on the net fluid flux (Pe1 + Pe2 ), which being constant for all three
simulations, fixes the top of hydrate layer to the same depth.
Figure 6.6 shows the variation of gas hydrate thickness and saturation as a
function of three different Pe1 , keeping the net fluid flux constant. It can be seen
that increasing Pe1 decreases the saturation within the fixed thickness of about two-
thirds of the GHSZ, but does not changes the vertical extent of hydrate occurrence.
109
The thickness of the gas hydrate layer only depends on the net fluid flux, Pe1 + Pe2 ,
and once this finite zone of hydrate occurrence is fixed by specifying the sum of
the Peclet numbers, hydrate saturation at any point within this zone depends on
how long this point resides within the GHSZ. The sedimentation rate governs this
residence time and, hence, controls the final average saturation. To summarize,
for geologic settings with only deeper methane sources, higher gas hydrate satu-
Predicting general conditions, like those formulated in the previous section, for bio-
genic in situ sources is not trivial. The effect of the normalized organic content
β is simple; increasing it always leads to more organic input and methane gen-
of the gas hydrate layer. Increasing the Damkohler number also causes more
methane generation within the GHSZ and yields higher saturations. The role of
As seen through the two examples of Blake Ridge and Peru Margin, a decrease
in opposite average gas hydrate saturation trends. This happens because Pe1 is
present in both the x-axis and y-axis terms of the contour plot (Figure 6.3), thereby
affecting the saturation in different ways. At lower Pe1 , when diffusive losses of
ever, at higher Pe1 diffusion is no longer a constraint and instead of being controlled
by Pe1 alone, the system now depends on the ratio Pe1 /Da. This ratio controls the
amount of methane generated from in situ sources within the GHSZ. Increase in
sedimentation rate now causes shorter residence times for the organic carbon and
hydrate phases in the GHSZ, resulting in lower average saturations. Hence, there
rates, respectively.
A similar argument can be formulated for gas hydrate systems with mixed sources
of methane, that is, both in situ biogenic as well as deeper sources. Methane from
an external deep source will only come into effect when |Pe2 | > |Pe1 |, otherwise the
net fluid flow will be downwards from the seafloor and the deeper methane will not
be advectively transported into the GHSZ. The net behavior of this mixed system
will have common elements of each individual system. We first study the effect of
the biogenic input parameters (Pe1 , Da, β) and then discuss the role of the external
Similar to the system where biogenic in situ methane is the only source, a mixed
source system will have higher gas hydrate saturations for larger values of Da and
β. The role of Pe1 , as mentioned before, is not so straightforward. Figure 6.7 shows
111
Normalized depth
0.6 0.6
0.8 0.8
1.0 1.0
1.2 1.2
1.4 1.4
1.6 1.6
1.8 1.8
2.0 2.0
0 0.02 0.04 0.06 0 0.1 0.2 0.3 0.4
Gas hydrate saturation Gas hydrate saturation
Figure 6.7: Variable effects of Pe1 on systems with mixed sources of methane.
Da = 5 and β = 3 for all simulations. (A) A diffusive mixed source system (Pe1 +
Pe2 < 1) shows that increasing Pe1 initially causes an increase in gas hydrate
saturation due to higher organic carbon flux into the sediments. Further, increase
in Pe1 results in a drop in hydrate saturation because sediment velocity become
high enough to cause large amounts of carbon pass through the GHSZ unreacted.
(B) For an advective mixed source system (Pe1 + Pe2 > 1) increasing Pe1 results in
a monotonous decrease in gas hydrate saturation due to shorter residence times
of the hydrate within the GHSZ.
the opposite effect of Pe1 on gas hydrate saturation for two cases: diffusion- and
advection-dominated systems. For example, when the net fluid flux is upward and
small in magnitude, so that diffusion dominates (Pe1 + Pe2 < 1), we observe that
Pe1 = 0.05 to Pe1 = 0.1 (Figure 6.7A). This happens because methane input to the
system from below the GHSZ is small at lower values of combined Peclet numbers
112
and any increase in Pe1 results in more organic input to the sediments, causing
higher hydrate saturation. Further increase in Pe1 causes the sediments to move
faster through the GHSZ causing lesser generation of methane within the stability
systems with mixed sources, on increasing Pe1 , hydrate saturation first increases,
At higher net Peclet numbers, fluid advection dominates the system and the
tation rate for this case results in smaller average gas hydrate saturation, because
now the main effect of higher Pe1 is to move the sediments faster through the
GHSZ (Figure 6.7B). This results in shorter residence times of the gas hydrate
Finally, the effect of varying Pe2 , with constant Pe1 , Da and β, is shown in Figure
6.8. The biogenic in situ generation of methane is fixed in this mixed source system
and only the upward flux of methane is changed. Pe2 for the three simulations in
Figure 6.8 was varied such that diffusion as well as advection dominated systems
are represented. It can be seen that increasing Pe2 leads to an increase in gas
hydrate saturation and makes the top of the gas hydrate layer shallower due to
greater recycling of methane back into the GHSZ. To summarize, for mixed source
systems, increasing Da, β and Pe2 always leads to higher gas hydrate saturations.
diffusion-dominated.
113
0
Pe1 + Pe2 = − 0.5
Pe + Pe = − 5.0
1 2
Pe1 + Pe2 = − 15.0
Normalized depth
0.5
1.0
1.5
0 0.05 0.10 0.15 0.20 0.25 0.30 0.35
Gas hydrate saturation
Figure 6.8: Effect of variable upward fluid flux Pe2 on gas hydrate saturation with
constant biogenic in situ methane generation. Pe1 , Da and β were set to 0.1,
5.0 and 3.0,respectively, for all three cases. Keeping biogenic input constant and
increasing the magnitude of net methane flux Pe1 + Pe2 (by increasing Pe2 ) causes
higher steady state gas hydrate saturations as well as shallower depths to the top
of the hydrate layer.
6.5 Conclusions
We have developed generalized average gas hydrate saturation contour plots that
sionless groups from our numerical model. We used one of these plots, which
assumes biogenic in situ generation as the only methane source, to model gas
hydrate provinces like Blake Ridge, Peru Margin and Costa Rica Margin. Applica-
tion of this contour plot to these three distinct settings shows good match between
114
predicted saturations and those inferred from proxy methods or other numerical
simulations. The saturation plot also serves as a simple tool to study the temporal
evolution of any gas hydrate system over geologic time-scales. This is achieved
by tracking the system’s path across the contours in response to the applied per-
sediment sequences at the Blake Ridge and Peru Margin, it is shown that the for-
mer system might be a depleting methane reservoir, whereas the latter might be
features, such as gas hydrate without free gas contact at steady state and higher
gas hydrate saturations at lower sedimentation rates, can also be explained by our
We also show that gas hydrate systems without free gas immediately below the
hydrate layer can exist in two distinct scenarios. At relatively lower TOC contents,
gas hydrate forms but remains isolated within the GHSZ, with no free gas formation
below. On increasing the TOC content, simulations show that free gas does form
beneath the base of the GHSZ, but remains disconnected from the bottom of the
hydrate layer. Further increase in TOC input causes both the hydrate and free gas
layers to approach the base of the GHSZ. The general contour plot shows that
average gas hydrate saturation within this special regime can achieve significant
amounts only at relatively low Pe1 , or low sedimentation rates. These features are
collectively used to model presence of hydrate and free gas, but without a BSR at
Site 994 (Blake Ridge) using an in situ methane source only. Contact between gas
hydrate and free gas at neighboring Sites 995/997 is then explained by using the
115
General conditions for high gas hydrate saturations are predicted for systems
dominated by either of the two sources of methane or a mixture of both. For hydrate
at high fluid flux and low sedimentation rates, with the thickness of the hydrate layer
only depending on the net fluid flux. In situ methanogenesis dominated settings
can be divided into diffusive and advective systems, with increase in sedimentation
rates causing higher hydrate saturations in the former regime. This behavior is
hydrate saturations for a wide variety of geologic settings, our model also provides
physical insight into the processes governing different hydrate systems. Although
detailed hydrate saturation profiles cannot be obtained from our contour plot, it
gives good estimates averaged over the regional scale without the need of any
Chapter 7
7.1 Introduction
Based on the supply of methane, marine gas hydrate systems can be distinguished
into two end-members: in situ systems where microbes generate methane within
the GHSZ (Claypool and Kvenvolden, 1983); and deep-source systems where ris-
ing fluids bring methane from depth (Hyndman and Davis, 1992). However, quan-
tifying gas hydrate abundance in either of the two systems remains a challenge,
For both gas hydrate systems, SMT denotes a relatively thin zone near the
seafloor where pore water sulfate and methane are depleted to zero (Figure 7.1).
Although microbes can also consume sulfate using solid organic carbon (Berner,
1980), AOM can dominate overall sulfate depletion in sediment sequences with
gas hydrates and modest methane fluxes (Borowski et al., 1996; Luff and Wallman,
117
Depth (mbsf)
80 Methane
15 solubility
100 L curve
t
Methane
120 20 Methane
140
25
160
180 Uf,ext 30
Figure 7.1: (A) Schematic representation of a gas hydrate system showing pore
water sulfate and methane concentrations, which go to zero at some shallow depth
because of anaerobic oxidation of methane (AOM). Also shown are methane sol-
ubility in water, the two fluid fluxes (Uf,sed and Uf,ext ), and depth to the base of
the gas hydrate stability zone (Lt ). (B) Close-up of the sulfate-methane transition
(SMT) showing overlap of sulfate and methane profiles, and its depth below the
seafloor (Ls ).
2003; Snyder et al., 2007). Further, since we focus only on deep-source systems
(i.e., sites with low organic carbon input), AOM becomes the only sulfate sink. The
sulfate profile and SMT depth in such systems should depend on methane flux from
below because of the simple 1:1 AOM reaction (Borowski et al., 1996). Additionally,
the vertical extent and abundance of gas hydrate is a function of upward methane
flux (Davie and Buffett, 2003a; Luff and Wallman, 2003). Thus, SMT depth (Ls ,
Figure 7.1) should relate to gas hydrate abundance, with relatively shallow SMTs
118
indicating elevated methane flux and high gas hydrate content (Borowski et al.,
1996, 1999; Davie and Buffett, 2003b; Luff and Wallman, 2003; Coffin et al., 2007).
relationship between methane flux and gas hydrate distribution (Bhatnagar et al.,
2007). Here, we expand that model by incorporating a SMT for deep source sys-
tems (Figure 7.1). We show that, at steady-state conditions, the depth of the SMT
relates to net fluid flux in the system and to average gas hydrate saturation (AGHS,
expressed as volume fraction of pore space) within the GHSZ. While previous stud-
ies have used sulfate profiles to constrain gas hydrate saturations at individual lo-
cations (Davie and Buffett, 2003b; Luff and Wallman, 2003), our model generalizes
the relationship between SMT depth and AGHS at any gas hydrate setting domi-
nated by methane flux from depth. This enables a first order estimation of AGHS
fluid flow, and external fluid flow (Bhatnagar et al., 2007). Figure 7.1 shows the
important system parameters governing these processes, along with the two key
depths in our model: Lt denoting the dimensional depth of the GHSZ, and Ls
representing the dimensional SMT depth below the seafloor. Following Bhatnagar
et al. (2007), the three-phase methane mass balance (liquid, gas hydrate and free
119
gas) can be written to include the AOM reaction in dimensionless form as:
" #
∂ 1 + γ φ̃ ³ ´
Sw c̃w
m + S c̃h
h m hρ̃ + S c̃g
g m gρ̃
∂ t̃ γ
· ¸
1+γ ∂ 1 + γ φ̃
+ (Pe1 + |Pe2 |) Ũf c̃w
m + Pe1 Ũs
³ ´ (Sh c̃hm ρ̃h + Sg c̃gm ρ̃g )
γ ∂ z̃ γ 1 − φ̃
" # Ã !
∂ 1 + γ φ̃ ∂c̃w MCH4 cw s,0 1 + γ φ̃
= Sw m − DaAOM Sw c̃w w
s c̃m (7.1)
∂ z̃ γ ∂ z̃ MSO4 cw
m,eqb γ
where Si represents saturation of phase i in pore space, ρ̃i is the density of phase i
and g denote liquid water, hydrate and free gas phases, respectively. The vertical
depth in our model is scaled by the depth of GHSZ (z̃ = z/Lt ). Consequently,
the SMT depth is also shown later in the results as a scaled depth L̃s = Ls /Lt .
(cw w
s ) is scaled by the seawater value (cs,0 ), to get the corresponding normalized
variables:
cim cws
c̃im = for i ∈ {w, h, g} , c̃w
s = (7.2)
cw
m,eqb cw
s,0
The reduced porosity parameters, φ̃ and γ, and normalized sediment flux, Ũs , are
defined as:
120
φ − φ∞ 1 − φ∞ Us
φ̃ = , γ= , Ũs = (7.3)
1 − φ∞ φ∞ Uf,sed
where φ is sediment porosity, φ∞ is the minimum porosity at great depth, and Uf,sed
is the fluid flux resulting from sedimentation and compaction. Porosity loss can
η φ0 − φ∞
φ̃ = , η= (7.4)
η + (1 − η) ez̃ 1 − φ∞
where η and φ0 are the reduced and actual porosities at the seafloor, respectively.
The fluid flux (Uf,sed ) can be related to the seafloor sedimentation rate (Ṡ) and
1 − φ0
Uf,sed = Ṡφ∞ (7.5)
1 − φ∞
The two Peclet numbers (Pe1 , Pe2 ) and the Damkohler number (DaAOM ) are defined
as:
Uf,sed Lt Uf,ext Lt ρw cw 2
m,eqb λLt
Pe1 = , Pe2 = , DaAOM = (7.6)
Dm Dm MCH4 Dm
where Uf,ext is the fluid flux due to external sources and λ is the second order rate
constant for AOM. Importantly, the first Peclet number (Pe1 ) characterizes the ratio
of sedimentation driven fluid flux to methane diffusion, while the second Peclet
number (Pe2 ) represents the ratio of external fluid flux to methane diffusion; the
121
" #
∂ 1 + γ φ̃ 1+γ ∂ h i
Sw c̃w
s + (Pe1 + |Pe2 |) Ũf c̃w
s =
∂ t̃ γ γ ∂ z̃
" # Ã !
∂ 1 + γ φ̃ Ds ∂c̃w
s 1 + γ φ̃
Sw − DaAOM Sw c̃w w
s c̃m (7.7)
∂ z̃ γ Dm ∂ z̃ γ
where Ds denotes sulfate diffusivity. The following initial and boundary conditions
I.C. : c̃w w
s (z̃, 0) = c̃m (z̃, 0) = 0 (7.8)
B.C.(1) : c̃w
s (0, t) = 1 , c̃w
m (0, t) = 0 (7.9)
∂c̃w
s cw
m,ext
B.C.(2) : (D, t) = 0 , c̃w w
m (D, t) = c̃m,ext = (7.10)
∂ z̃ cw
m,eqb
where cw
m,ext is the methane concentration in the external flux and D denotes the
7.3 Results
Equations (7.1) and (7.7) are solved numerically to obtain steady-state profiles for
methane, gas hydrate saturation, and pore-water sulfate. For results shown later,
(Bhatnagar et al., 2007). The AGHS, however, is more sensitive to seafloor depth.
7.2 and 7.3) use a seafloor depth of 1000 m, whereas the general relationship be-
tween AGHS and SMT depth (Figure 7.4) summarizes results for multiple seafloor
depths. Porosity at seafloor (φ0 ) and at depth (φ∞ ) are assumed to be 0.7 and 0.1,
m2 /s, respectively (Iversen and Jørgensen, 1993), chm is set to 0.134, seawater sul-
fate concentration equals 28 mM, and ρh and ρf are assumed to be 930 and 1030
that c̃w
m,ext equals unity.
We first study the effect of DaAOM on steady-state profiles. For fixed Pe1 and
Pe2 , decreasing DaAOM results in a relatively thick SMT zone (Figure 7.2a). Higher
causing a relatively sharp SMT. The SMT is usually less than a few meters at most
gas hydrate settings (Borowski et al., 1999; Riedel et al., 2006), so we use a large
Concentration profiles simulated for three different sets of Pe1 and Pe2 , but with
the sum Pe1 + Pe2 held constant at -10, are shown in Figure 7.2b. Overlap of these
profiles demonstrates that neither Pe1 nor Pe2 individually controls the system, but
that their sum determines the concentrations and the scaled SMT depth, L̃s . This
123
0.2
0.02 0.04
Normalized depth ( z~ = z/L )
SMT for 8
t
0.06 0.12
0.8 <S >=2%
0.07 0.14 h
Methane
0.08 0.16
1.0
Pe =0.1,Pe =−10.1
1 2 <Sh>=0.5%
0.09 0.18 Pe =0.4,Pe =−10.4
1 2
Pe =1.0,Pe =−11.0
1 2 <Sh>=0.2%
0.10 0.20 1.2
0 0.5 1 0 0.5 1 0 5 10
Normalized concentration Normalized concentration Hydrate saturation
(% pore vol.)
Figure 7.2: Effect of Damkohler number (DaAOM ) and Peclet numbers (Pe1 , Pe2 )
on steady state profiles. Pe1 + Pe2 = −10 for all simulations (Note different y-axis
scale for each plot). (a) Sulfate and methane profiles for DaAOM = 108 (solid curves)
and DaAOM = 106 (dashed curves). The hatched regions compare the thickness
of the SMT zone for the two cases. (b) Simulations for different sets of Pe1 and
Pe2 , but with Pe1 + Pe2 = −10. Overlap of methane and sulfate profiles shows that
the sum Pe1 + Pe2 controls the concentrations. (c) Gas hydrate saturation profiles
do not depend only on Pe1 + Pe2 , because different Pe1 causes distinct residence
times of gas hydrate within the GHSZ.
124
sum, Pe1 + Pe2 , represents the net fluid flux through the system. Hydrate saturation
profiles, however, depend on more than the sum of the Peclet numbers (Figure
7.2c). The AGHS, denoted hSh i, for each of the three cases, is about 0.2%, 0.5%
and 2%, with the highest value corresponding to the smallest Pe1 (0.1) and largest
Pe2 (-10.1). Small Pe1 and large Pe2 correspond to low sedimentation rate and high
methane flux, respectively. High external flux and low sedimentation rate result
in higher methane input to the system and longer residence time for gas hydrate
within the GHSZ, causing higher AGHS. However, for all three cases, the product
Pe1 hSh i is the same. Thus, Figure 7.2c demonstrates that this product (Pe1 hSh i),
which characterizes the flux of gas hydrate through the GHSZ, is only a function of
Increasing net methane flux from depth (i.e., raising the magnitude of Pe1 + Pe2 )
results in a shallow scaled SMT depth (Figure 7.3a), as proposed by Borowski et al.
(1996, 1999). Increasing Pe2 , with Pe1 held constant, increases gas hydrate sat-
uration (Figure 7.3b) due to higher methane input to the system. Consequently,
the product Pe1 hSh i also increases. Hence, the scaled depth to the SMT, L̃s , and
the product Pe1 hSh i both depend on the sum Pe1 + Pe2 . As a consequence, scaled
SMT depth and Pe1 hSh i become correlated (Figure 7.4). Thus, AGHS can be esti-
mated using Figure 7.4 if L̃s and Pe1 are known from site data.
125
0 0
Pe1+Pe2 = − 5
0.02 Sulfate Pe +Pe = − 10
1 2
0.2 Pe +Pe = − 15
1 2
Normalized depth (z~ = z/L ) 0.04
0.06
0.4
Methane
0.08
0.10 0.6
0.12
0.8
0.14
0.16 Pe1+Pe2 = − 5
Pe +Pe = − 10 1.0
1 2
0.18 Pe +Pe = − 15
1 2
0.20 1.2
0 0.5 1 0 5 10 15
Normalized concentration Hydrate saturation (% pore vol.)
Figure 7.3: Effect of net fluid flux (Pe1 + Pe2 ) on steady-state concentrations. Pe1
equals 0.1 for all simulations. (a) High Pe1 + Pe2 defines higher net methane fluxes,
resulting in shallower SMT zones. (b) Gas hydrate saturation at steady state in-
creases as Pe1 + Pe2 increases.
Sites drilled by Ocean Drilling Program (ODP) Leg 146 and Integrated Ocean
Drilling Program Expedition (IODP) 311 penetrate gas hydrate systems along Cas-
cadia Margin (Westbrook et al., 1994; Riedel et al., 2006). The low organic carbon
content of sediment and pervasive upward fluid migration at these sites suggests
that gas hydrate in the region is controlled by methane supplied from depth (Riedel
et al., 2006). The procedure to calculate AGHS, hSh i, from site-specific data is now
summarized:
126
0.4
0.3
Site U1325
0.2
Site 889
0.1
0
0 0.01 0.02 0.03 0.04 0.05 0.06
~
Scaled SMT depth: Ls
Figure 7.4: Relationship between Pe1 hSh i and scaled SMT depth (L̃s = Ls /Lt ) for
several seafloor depths. Points corresponding to different Cascadia Margin sites
are plotted to show how AGHS can be estimated from L̃s using this plot (Table 7.1).
• Use sedimentation rate (Ṡ) to calculate Pe1 from equations (7.5) and (7.6);
• Calculate the normalized SMT depth L̃s using the dimensional depths Ls and
Lt ;
• For given seafloor depth and L̃s , obtain average hydrate flux Pe1 hSh i from
Figure 7.4;
These calculations and comparison with estimated AGHS are shown in Table 7.1.
127
Site 889 (ODP Leg 146) has been previously modeled as a gas hydrate system
dominated by deeper methane sources, with the pore water chloride profile indicat-
ing a peak hydrate saturation between 1.5-2% near the base of GHSZ, and AGHS
<1% within the GHSZ (Davie and Buffett, 2003a). Although several fundamental
problems confront such estimates (Egeberg and Dickens, 1999), this result agrees
fairly well with our calculated AGHS of 0.4% across the GHSZ, using the SMT as
For the IODP Expedition 311 sites, we compute gas hydrate saturation profiles
and AGHS using LWD resistivity log data and Archie parameters given in Riedel
et al. (2006) (Table 7.1). Overall, hydrate saturations derived using resistivity logs
and those simulated using our model match well, although simulations consistently
128
have lower AGHS than the log predictions at all four IODP sites. A possible expla-
nation for this deviation is that interpretations of resistivity logs depend on knowl-
edge of formation water resistivity and three empirical constants (Archie, 1942),
predictions have gas hydrate first occurring well below the seafloor (Figures 7.2
and 7.3), compared to log-based results that often predict gas hydrate immedi-
ately below the seafloor. Apart from the small deviations between model and log
predictions, and more importantly, our model captures the variation in average sat-
7.5 Conclusions
We show that scaled depth to the SMT can be used to estimate AGHS for deep
source gas hydrate systems. Simulation results demonstrate that the net fluid flux
uniquely determines the scaled depth to SMT (L̃s ) as well as the average gas hy-
drate flux (Pe1 hSh i) through the GHSZ. Results also show that shallow scaled SMT
depths and low Peclet number (Pe1 ) lead to higher hydrate saturations. Application
of this method to sites along Cascadia Margin reveals a good match with observed
hydrate saturations and accurately predicts lateral variability in gas hydrate satura-
tion.
129
Chapter 8
8.1 Introduction
energy resource (Collett, 2002; Buffett and Archer, 2004; Milkov, 2004), a natural
geohazard (Sultan et al., 2004), a component of the global carbon cycle (Dickens,
2003) or an agent of climate change (Dickens et al., 1995; Kennett et al., 2003).
that make use of some anomaly associated with presence of gas hydrates in the
sediment (Paull et al., 2000; Hesse, 2003; Tréhu et al., 2004). Pore water anoma-
lies, in particular, have been used in conjunction with coupled transport models
often include dissolved halides (usually chloride), ammonium and sulfate (Ege-
berg and Barth, 1999; Ussler and Paull, 2001; Davie and Buffett, 2003a,b; Haeckel
et al., 2004; Torres et al., 2004; Hensen and Wallman, 2005; Wallman et al., 2006).
Although chloride anomalies seem to be the most widely used proxy in numerical
line profile (Egeberg and Dickens, 1999). Different choices of this baseline chloride
profile will lead to different gas hydrate saturations (Ussler and Paull, 2001). Fur-
ther, gas hydrate dissociation at the base of the GHSZ causes localized freshening
of the pore water, which when brought back into the GHSZ by upward fluxes, can
result in a decrease in the chloride anomaly (Ussler and Paull, 2001; Hesse, 2003).
Finally, in geologic settings where only shallow piston cores have been collected,
gas hydrate saturation cannot be predicted for the depth interval below the cored
section.
ter sulfate profile as the main proxy for gas hydrate saturation. This is motivated
by the fact that most of the geologic settings where gas hydrate has been recov-
Tréhu et al., 2004; Riedel et al., 2006). The advantage of this approach is that
any baseline curves. Moreover, our analysis differs from the traditional approach
that involves simulating gas hydrate distribution using numerical models and then
matching them with observed sulfate profiles at a given site (Davie and Buffett,
2003b; Luff and Wallman, 2003; Hensen and Wallman, 2005). We follow the in-
verse route and predict gas hydrate saturation directly from the sulfate profile.
SMT in gas hydrate simulations (Bhatnagar et al., 2008). It was shown that the
net upward fluid flux in the system controls the pore water methane and sulfate
131
concentration profiles, and thus the depth to the SMT, as well as the average gas
hydrate flux through the GHSZ. We simulated several different parameter sets to
generalize this relationship, but the simulation methodology suffers from a few dis-
advantages. First, the entire gas hydrate saturation profile cannot be constructed
from the general relationship between SMT depth and gas hydrate flux. New sim-
ulations have to be performed to get the saturation profile, which can get compu-
tationally expensive in order to resolve the thin SRZ. Second, the relation between
gas hydrate flux and SMT depth was generalized in our numerical study for certain
ical expressions, exact solutions can be obtained for the site-specific parameters
main reactions. In the absence of methane in the system, sulfate gets reduced
1980; Boudreau and Westrich, 1984). The other important sink of sulfate in these
systems is the anaerobic oxidation of methane (AOM), which uses dissolved methane
for sulfate reduction through the following reaction (Borowski et al., 1996):
CH4 + SO2−
4 −→ HCO− −
3 + HS + H2 O (8.1)
132
Depth (mbsf)
curve
Methane L −L
15 t h
Lt
100
Lh Methane
20
150 25
U 30
f,ext
flux towards the seafloor, which can make this second route for sulfate depletion
significant (Borowski et al., 1996; Niewöhner et al., 1998; Davie and Buffett, 2003b;
Modeling results of Davie and Buffett (2003b) applied to Site 997 of the Blake
Ridge region suggest that sulfate depletion is mainly governed by the AOM reac-
tion in the presence of deeper dissolved methane. Recently, Snyder et al. (2007)
flux, in accordance with the 1:1 ratio suggested by the AOM reaction, across three
sites along Umitaka Spur, Japan Sea. Dominance of the AOM reaction over the
organic carbon pathway in depleting sulfate is also suggested by the linearity and
which cannot be described by the variability of sedimentation rates and TOC con-
tent at these sites (Borowski et al., 1996, 1999). The Blake Ridge is characterized
by modest TOC content (∼1.5 %, (Paull et al., 2000)). Hence, if sulfate depletion
profile at these sites, we can safely assume the AOM reaction to be the only sul-
fate sink in our model. This is possible because, as mentioned before, we focus
on hydrate settings with low TOC where all methane is supplied by deeper sources
only.
134
In this section we first derive a relationship between sulfate flux and depth of the
SMT zone using a steady-state sulfate mass balance. This also enables calculation
of the sulfate concentration profile as a function of vertical depth below the seafloor.
We then use the equality of molar fluxes of sulfate and methane at the SMT to
express the results in terms of the methane flux from depth. This is followed by
writing a two-phase methane balance for the system, which links the thickness
of the gas hydrate layer and gas hydrate saturation to the methane flux. Finally,
by eliminating methane flux between the sulfate and methane mass balances, we
show how the SMT depth is related to thickness of the gas hydrate layer and its
saturation. All equations are converted to dimensionless form to help reduce the
The methane flux input to the system from depth would depend on the net fluid
flux as well as the methane concentration in the rising pore waters. In this formu-
lation we assume that at steady state the gas hydrate layer extends to the base of
concentration in the pore fluid at this depth is set equal to the peak solubility value
We start with the sulfate mass balance applied between the seafloor and the SMT.
carbon;
2. Both methane and sulfate react at the SMT fast enough such that their con-
Geochemical data at most gas hydrate settings suggests that a finite region of
later, we normalize the vertical depth in our model by the depth to the base of the
GHSZ. This causes the finite region of sulfate-methane overlap, which is usually
a few meters thick, to become thin in the dimensionless form and approach a
We start with the steady-state sulfate mass balance, which is written as:
" #
d dcl
Uf,tot ρf cls − φρf Ds s = 0, 0 < z < Ls (8.2)
dz dz
where Uf,tot is the net fluid flux, ρf is pore water density, cls is the mass fraction of
sulfate in pore water, φ denotes porosity, Ds is sulfate diffusivity and Ls is the depth
to the SMT (Figure 8.1). The vertical depth z is set to zero at the seafloor and is
positive downwards. The mass balance (8.2) implies that the mass flux of sulfate,
FSO4 , remains constant within the SRZ, and can be rewritten as:
dcls
Uf,tot ρf cls − φρf Ds = FSO4 , 0 < z < Ls (8.3)
dz
136
This equation is now expressed in dimensionless form. The vertical depth scale
is normalized by Lt , the depth to the base of the GHSZ (z̃ = z/Lt , L̃s = Ls /Lt ),
while sulfate concentration is scaled by its seawater value, c0SO4 , to rewrite the
mass balance in terms of the normalized concentration (c̃ls = cls /c0SO4 ). The total
fluid flux is written as the sum of its two components: Uf,sed due to sedimentation-
compaction and Uf,ext due to upward external sources (Appendix A1). This enables
us to define two separate Peclet numbers that compare each fluid flux to methane
diffusion, as follows:
Uf,sed Lt Uf,ext Lt
Pe1 = , Pe2 = (8.4)
Dm Dm
à ! à ! à !
1+γ 1 + γ φ̃ Ds dc̃ls 1 FSO4 Lt
(Pe1 + Pe2 ) c̃ls − =
γ γ Dm dz̃ 1 − φ∞ ρf c0SO4 Dm
, 0 < z̃ < L̃s (8.5)
1−φ∞
where φ̃ is the reduced porosity, γ is φ∞
, and φ∞ is the minimum porosity achieved
at great depth. The porosity model, assuming hydrostatic pore pressure and equi-
à !
1+γ
(Pe1 + Pe2 ) = Q (8.6)
γ
137
à !
1 FSO4 Lt
= fSO4 (8.7)
1 − φ∞ ρf c0SO4 Dm
à !
1 + γ φ̃ dc̃l
Qc̃ls − D̃s s = fSO4 (8.8)
γ dz̃
where D̃s is the ratio of sulfate to methane diffusivity. Two boundary conditions
are used to get an analytical relationship between sulfate flux and the SMT depth.
The first one is applied at the seafloor where the normalized sulfate concentration
is equal to unity, while the second is applied at the base of the SRZ, where both
Z 0Ã ! Z 1 Ã !
γ D̃s
dz̃ = dc̃ls (8.11)
L̃s 1 + γ φ̃ 0 Qc̃ls − fSO4
à !
D̃s Q
g(z̃)|z̃=0 − g(z̃)|z̃=L̃s = ln 1 − (8.12)
Q fSO4
138
where g(z̃) represents the integral of the porosity term on the left hand side of
³ ´
γ z̃ + γ 2 ln η(1 + γ) + (1 − η)ez̃
g(z̃) = (8.13)
1+γ
The above relation is a non-linear function of the scaled depth z̃. However, for the
case of no compaction, g(z̃) can be expressed as a simple linear function of z̃. This
g(z̃) = γ z̃ (8.14)
The function g(z̃) is plotted in Figure 8.2 for different values of the reduced
ular depth z̃ simply as g[z̃], so that g(z̃)|z̃=0 and g(z̃)|z̃=L̃s are written as g[0] and
Q
fSO4 = ³
Q
h i´ (8.15)
1 − exp D̃s
g[0] − g[L̃s ]
grated from [z̃, L̃s ], which yields a relation similar to equation (8.12), and is written
as:
à !
D̃s Qc̃ls (z̃)
g[z̃] − g[L̃s ] = ln 1 − (8.16)
Q fSO4
139
0
No compaction (η=0), γ = 9
0.1 η = 2/9, γ = 9
η = 6/9, γ = 9
Normalized depth ~
z 0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20 25
~
Function g(z)
Figure 8.2: The function g(z̃), equation(8.13), plotted for three different porosity
parameters. The dashed line represents the case of zero compaction η = 0, the
dash-dot curve represents the case η = 2/9 (or φ0 = 0.3), while the solid curve
denotes the case η = 6/9 (or φ0 = 0.7). The parameter γ = 9 (or φ∞ = 0.1) for all
cases.
à !
Qc̃ls (z̃) Q ³ ´
= 1 − exp g[z̃] − g[L̃s ] , 0 < z̃ < L̃s (8.17)
fSO4 D̃s
Finally, substituting the expression for fSO4 (equation (8.15)) into the above equa-
h ³ ´i
Q
1 − exp D̃s
g[z̃] − g[L̃s ]
c̃s (z̃) = h ³ ´i , 0 < z̃ < L̃s (8.18)
1− exp D̃Qs g[0] − g[L̃s ]
140
depth, L̃s , and the net fluid flux, Q. We later show that L̃s depends uniquely on Q,
which helps us to obtain the sulfate concentration profile by just specifying L̃s at
a given hydrate site. Thus, sulfate concentration profiles from equation (8.18) for
different values of SMT depth (L̃s ) are shown later in the Results section.
We now rewrite equation (8.15) in terms of the methane flux from depth. At the
base of the SRZ, the magnitude of the molar fluxes of methane and sulfate are
equal due to the 1:1 stoichiometry of the AOM reaction. Thus, the sulfate mass
flux (FSO4 ) can be written in terms of the methane mass flux (FCH4 ) from below as
follows:
MSO4
FSO4 = − FCH4 at z̃ = L̃s (8.19)
MCH4
(8.7) yields:
à !
1 1 MSO4 Lt
fSO4 =− FCH4 (8.20)
1 − φ∞ ρf c0SO4 MCH4 Dm
à !
1 FCH4 Lt
fCH4 = (8.21)
1 − φ∞ ρf clm,eqb Dm
where clm,eqb is the methane solubility at the base of GHSZ. Using this notation,
equation (8.20) is used to express the dimensionless methane flux in terms of the
MSO4 clm,eqb
fCH4 = −fSO4 /m , where m= (8.22)
MCH4 c0SO4
SMT depth (L̃s ) and methane flux (fCH4 ), which is rewritten below using equations
Q/m
fCH4 = −fSO4 /m = h
Q
³ ´i (8.23)
1 − exp D̃s
g[0] − g[L̃s ]
We now perform mass balances on methane and water and apply them to two
distinct spatial domains. The first domain extends from the SMT to the top of the
gas hydrate layer, whereas the second domain extends from the top of the gas
hydrate layer to the base of the GHSZ (Figure 8.1). These two mass balances,
coupled with the sulfate balance in the previous section, are used to solve the
balance, valid from the SMT depth to the base of the GHSZ:
· ¸
∂ Us ∂cl
Uf clm ρf + φSh chm ρh − φ(1 − Sh )ρf Dm m = 0
∂z 1−φ ∂z
, L s < z < Lt (8.24)
where Us denotes sediment flux, Sh denotes gas hydrate saturation (volume frac-
tion of pore space), and chm is the methane mass fraction in the hydrate phase (a
constant, chm = 0.134 for structure I hydrate (Sloan and Koh, 2007). This methane
Us ∂cl
Uf clm ρf + φSh chm ρh − φ(1 − Sh )ρf Dm m = FCH4 , L s < z < Lt (8.25)
1−φ ∂z
the above equation, we scale sediment flux by Uf,sed , methane mass fractions by
methane solubility at the base of the GHSZ (clm,eqb ) and gas hydrate density by the
Us cl ch ρh
Ũs = , c̃lm = l m , c̃hm = l m , ρ̃h = (8.26)
Uf,sed cm,eqb cm,eqb ρf
Using the water mass balance, the methane mass balance is written in the follow-
143
à !à ! à !
Pe1 Ũs 1+γ 1 + γ φ̃ ³ ´ 1 + γ φ̃ ∂c̃l
Qc̃lm + Sh ρ̃h c̃hm − chw c̃lm − (1 − Sh ) m
1 − φ̃ γ γ γ ∂ z̃
= fCH4 , L̃s < z̃ < 1 (8.27)
where the terms Q and fCH4 have been defined previously (equations (8.6) and
(8.21)).
The methane mass balance (8.27) is first applied to the region extending from the
SMT to the top of the hydrate layer. Let the thickness of the gas hydrate layer be
Lh , which implies that depth from the seafloor to the top of the gas hydrate layer
L̃h = Lh /Lt , while depth to the top of hydrate is (1 − L̃h ). This region (z < 1 − L̃h )
does not contain any hydrate, so that equation (8.27) can be simplified by setting
Sh = 0:
à !
1 + γ φ̃ ∂c̃lm
Qc̃lm − = fCH4 , L̃s < z̃ < 1 − L̃h (8.28)
γ ∂ z̃
Methane concentration in this region (L̃s < z̃ < 1 − L̃h ) is bounded by zero at the
depth of SMT (L̃s ) and by the methane solubility curve at the top of the gas hydrate
144
layer. Hence, the two boundary conditions for this equation are:
where c̃m,sol (z̃) = cm,sol (z̃)/clm,eqb is the normalized methane solubility curve within
the GHSZ (i.e., methane mass fraction in pore water in equilibrium with gas hy-
drate) as a function of the scaled depth z̃. Similar to the function g[z̃], we denote
function evaluations of the solubility curve as c̃m,sol [z̃], so that c̃m,sol (z̃)|z̃=1−L̃h is
simply denoted as c̃m,sol [1 − L̃h ]. Analogous to the solution of the sulfate balance,
equation (8.28) can be integrated with the above boundary conditions as:
Using the function g[z̃] defined in equation (8.13), the above equation is written as:
à !
1 Qc̃m,sol [1 − L̃h ]
g[1 − L̃h ] − g[L̃s ] = ln 1 − (8.32)
Q fCH4
which can be rearranged to yield the methane flux in terms of the two scaled depths
manner similar to the sulfate concentration profile discussed in section 8.3.1. The
145
h ³ ´i
1 − exp Q g[z̃] − g[L̃s ]
c̃m (z̃) = c̃m,sol [1 − L̃h ] h ³ ´i , L̃s < z̃ < 1 − L̃h (8.34)
1 − exp Q g[1 − L̃h ] − g[L̃s ]
The general methane mass balance equation (8.27) is now applied to the region
extending from the top of hydrate layer to the base of the GHSZ. In this region, the
gas hydrate saturation (Sh ) becomes the primary dependent variable. Substituting
c̃m,sol [z̃] for the pore-water concentration c̃lm (z̃) into equation (8.27), we get the
à !à !
Pe1 Ũs 1 + γ 1 + γ φ̃
Qc̃m,sol [z̃] +
1 − φ̃ γ γ
à !
³ ´ 1 + γ φ̃ 0
Sh ρ̃h c̃hm − chw c̃m,sol [z̃] − (1 − Sh )c̃m,sol [z̃]
γ
= fCH4 , 1 − L̃h < z̃ < 1 (8.35)
0
where c̃m,sol [z̃] denotes the derivative of the solubility curve at any given depth z̃.
Several previous simulation studies have shown that gas hydrate saturation
monotonously increases from zero at the top of the gas hydrate layer to a maximum
value at the base of the GHSZ (Davie and Buffett, 2001, 2003a; Bhatnagar et al.,
146
2007, 2008). We use this observation to impose the constraint that gas hydrate
saturation goes to zero as the top of the hydrate layer is approached. This condition
Sh → 0 as z̃ → (1 − L̃h )+ (8.36)
à !
1 + γ φ̃ 0
Qc̃m,sol [z̃] − c̃m,sol [z̃] = fCH4 , z̃ → (1 − L̃h )+ (8.37)
γ
We now have three equations (8.23, 8.33, 8.37) in terms of four unknowns (L̃s ,
fCH4 , Q, and L̃h ). Hence, by using scaled SMT depth (L̃s ) as an input from site
data, the other three unknowns can be calculated. The next section illustrates
this procedure, whereas the complete steps to perform the overall calculation are
In this section, we obtain two non-linear coupled equations in terms of the three
variables L̃s , L̃h , and Q. This is achieved by eliminating fCH4 from the three mass
balance equations (8.23, 8.33, 8.37). First, we eliminate fCH4 between equations
(8.23) and (8.33), which amounts to equating the sulfate flux to the methane flux
the methane flux in the region containing gas hydrate. This helps to eliminate fCH4
à !
Qc̃m,sol [1 − L̃h ] 1 + γ φ̃ 0
h ³ ´i = Qc̃m,sol [1 − L̃h ] − c̃m,sol [1 − L̃h ](8.39)
1 − exp Q g[1 − L̃h ] − g[L̃s ] γ
Thus, once L̃s is known for a particular site, equations (8.38) and (8.39) can be
the scaled thickness L̃h and modified sum of Peclet numbers (Q). Apart from L̃s ,
the site-specific parameters needed to completely specify the system include the
minimum and maximum reduced porosities (γ and η, Appendix A1), diffusivity ratio
(D̃s ), parameter m (eq. (8.22)), and the methane solubility curve within the GHSZ
(c̃m,sol [z̃]).
It should be noted that the Peclet numbers occur in equations (8.38) and (8.39)
as the modified sum, Q, rather than Pe1 or Pe2 . This implies that we do not need
or L̃h . However, to compute the gas hydrate saturation profile, Pe1 needs to be
the steady-state gas hydrate saturation profile below the top of the hydrate layer
à !à !
Pe1 Ũs 1+γ 1 + γ φ̃ ³ ´
Sh ρ̃h c̃hm − chw c̃m,sol [z̃] −
1 − φ̃ γ γ
à !
1 + γ φ̃ 0
(1 − Sh )c̃m,sol [z̃] = fCH4 − Qc̃m,sol [z̃] , 1 − L̃h < z̃ < 1 (8.40)
γ
fCH4³−Qc̃m,sol
´ [z̃] + c̃
0
As mentioned in the previous section, specifying L̃s allows calculation of L̃h and
Q through solution of the coupled equations (8.38) and (8.39). Substituting these
variables into any of the methane flux expressions (for example, equation (8.33))
yields the methane flux fCH4 . Using these values of L̃h , Q, fCH4 , and other sys-
tem parameters, equation (8.41) gives the complete gas hydrate saturation profile
We have shown through numerical simulations that the steady state gas hydrate
flux through the GHSZ is related to the scaled SMT depth, through their depen-
dence on the net fluid flux in the system (Bhatnagar et al., 2007). The average
gas hydrate flux was defined as the product of the Peclet number , Pe1 , and gas
hydrate saturation averaged over the entire GHSZ, hSh i. This relationship helps
in the estimation of hSh i from L̃s (Bhatnagar et al., 2007). We now show that this
The first term in the denominator of equation (8.41) contains the expression
³ ´
c̃hm − chw c̃m,sol [z̃] , where c̃hm was defined as c̃hm = chm /clm,eqb , which for most marine
systems is of the order of c̃hm = 0.134/10−3 ≈ 102 . The other two terms chw c̃m,sol [z̃]
0
and c̃m,sol [z̃] are usually less than unity. This implies that c̃hm will be about two orders
in magnitude greater than the other terms in the denominator and this approxima-
fCH4 −Qc̃m,sol [z̃]
³ ´ + c̃0
1+γ φ̃ m,sol [z̃]
γ
Sh ≈ Pe1 Ũs
³
1+γ
´ , 1 − L̃h < z̃ < 1 (8.42)
1−φ̃ γ
ρ̃h c̃hm
Finally, the above equation can be written in terms of the product Pe1 Sh :
fCH4 −Qc̃m,sol [z̃]
³ ´ + c̃0
1+γ φ̃ m,sol [z̃]
γ
Pe1 Sh ≈ Ũs
³
1+γ
´ , 1 − L̃h < z̃ < 1 (8.43)
1−φ̃ γ
ρ̃h c̃hm
Equation (8.43) can be integrated over depth to give the term Pe1 hSh i as a function
150
fCH4 −Qc̃m,sol [z̃]
³ ´ + c̃0
Z 1 1+γ φ̃ m,sol [z̃]
γ
Pe1 hSh i ≈ Ũs
³
1+γ
´ dz̃ (8.44)
1−L̃h
1−φ̃ γ
ρ̃h c̃hm
One of the assumptions in the previous analysis is that gas hydrate exists within
the GHSZ, and at steady-state extends to the base of the GHSZ. However, if the
upward methane flux is low, pore-water methane concentration might not exceed
the local solubility and will result in no hydrate formation. This situation is de-
picted schematically in Figure 8.3, which compares methane and sulfate concen-
tration profiles for two distinct methane fluxes. For relatively low fluxes (dashed set
of curves), methane concentration does not exceed the local solubility anywhere
within the GHSZ, causing no hydrate formation. A finite SMT still exists for this
case. Progressively increasing the methane flux from depth will cause methane
concentration to approach the solubility curve and for a certain value of this flux,
the methane concentration becomes equal to the solubility curve at some depth
within the GHSZ. This case, depicted by the solid set of curves, corresponds to the
minimum methane flux required for gas hydrates to form. Any methane flux greater
than this minimum will have a SMT as well as finite gas hydrate saturation within
the GHSZ.
a minimum flux required to form hydrates, with the actual value being dependent on
151
(a) (b)
0 0
Sulfate
Methane
solubility 0.05
curve Maximum
0.5 Methane
SMT depth
for hydrates
to form
Normalized depth
Normalized depth
0.10
1.0
Minimum thickness SMT with
of gas hydrate layer no hydrate
(higher flux)
0.15 formation
1.5
No hydrate
0.20
formation
(lower flux)
2.0
0.25
2.5
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0
Normalized concentration Normalized concentration
Figure 8.3: (a) Schematic representation of a gas hydrate system showing effect of
upward fluid fluxes on pore water sulfate and methane concentration profiles. The
solid curves represent the minimum flux case for which the methane concentration
just exceeds the local solubility curve, causing hydrates to form. Any fluid flux lower
than this minimum value (dashed curves) will not be able to exceed the solubility
curve or form any gas hydrate, though a relatively deeper SMT will still exist. (b)
Close-up of the same plot showing the maximum allowed SMT depth (solid curves)
for a given gas hydrate system. The deeper SMT depth (dashed curves) exceeds
the maximum allowed value, implying no gas hydrate formation for this case.
152
the solubility curve and the methane concentration in the rising fluids. This feature
indirectly imposes a maximum limit on the SMT depth (for hydrates to precipitate)
as well as a minimum thickness of the gas hydrate layer, L̃h,min . We now derive
Gas hydrates start to precipitate when the methane concentration curve just
requiring the two curves to become tangential to each other. Let this tangency
occur at depth z̃ = 1 − L̃h,min , which marks the minimum thickness of the gas
Using expressions for concentration profiles and fluxes derived earlier, the above
conditions can be recast into the following set of non-linear equations (see Ap-
" Ã ! #
1 + mc̃m,ext g[1 − L̃h,min ] − g[L̃s,max ]
1 − exp D̃s ln
mc̃m,ext g[0] − g[L̃s,max ]
c̃m,sol [1 − L̃h,min ]
− =0 (8.47)
c̃m,ext
and
153
³ ´ Ã !
1+mc̃m,ext 0
·
D̃s ln mc̃m,ext
g [1 − L̃h,min ] 1 + mc̃m,ext
exp D̃s ln
g[0] − g[L̃s,max ] mc̃m,ext
¸ 0
g[1 − L̃h,min ] − g[L̃s,max ] c̃ [1 − L̃h,min ]
+ m,sol =0 (8.48)
g[0] − g[L̃s,max ] c̃m,ext
0
where g represents the derivative of the function g[z̃], the depth L̃s,max corresponds
to the maximum SMT depth at tangency, and c̃m,ext is the normalized methane
The above set of equations contain two unknowns, L̃h,min and L̃s,max , in terms
of the two system parameters, c̃m,sol [z̃] and c̃m,ext . For simplicity, we henceforth
assume that pore water from depth is saturated with methane so that c̃m,ext = 1.
Thus, any given gas hydrate setting, characterized by its unique solubility curve
through local seafloor depth, temperature and geotherm conditions, will have a
unique maximum L̃s (= L̃s,max ) and minimum L̃h (= L̃h,min ) required for gas hy-
drate to be present. Hence, before actual calculations for gas hydrate saturation
are performed, the observed SMT depth should be checked against the maximum
allowed value to see whether hydrates exist or not. Values of L̃s,max and L̃h,min for
8.4 Results
1. Given the local solubility curve c̃m,sol [z̃] and other site-specific parameters
154
(m, D̃s , γ and η), solve equations (8.47) and (8.48) to obtain the maximum
allowed SMT depth L̃s,max . If the observed scaled SMT depth, L̃s , for this site
2. If L̃s < L̃s,max , solve coupled equations (8.38) and (8.39) to obtain the modi-
3. Using these values, calculate methane flux, fCH4 , from any of the the three
4. Substitute into equations (8.18) and (8.34) to get the sulfate and methane
5. By specifying the parameters ρ̃h , Pe1 , c̃hm and chw , equation (8.41) gives the
the GHSZ. It can be calculated for any geologic setting either through rigorous
tionships (Davie et al., 2004; Tishchenko et al., 2005). A simple exponential type
dependence of this solubility on depth has been proposed by Davie et al. (2004).
For sake of demonstration and simplicity, we also approximate the solubility curve
where r1 and r2 are fitting constants. As mentioned before, this solubility curve is
scaled by methane solubility at the base of the GHSZ, so that its normalized value
r1 and r2 :
This simple equation, with a single fitting parameter r2 , yields very good fits to solu-
bility curves (Figure 8.4) obtained through rigorous thermodynamic models (Bhat-
nagar et al., 2007), for two different seafloor depths, seawater salinity, seafloor
We now explain the system in terms of the input parameter, L̃s . The following
constant parameter values are assumed for all results shown later: η = 6/9, γ = 9
(which correspond to φ0 = 0.7, φ∞ = 0.1), chm = 0.134, ρ̃h = 0.9, MCH4 = 16 g/mol,
MSO4 = 96 g/mol, seawater sulfate concentration equals 28 mM, and D̃s = 0.64
156
0.2
0.3
Normalized depth
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 0.5 1.0 1.5
Normalized methane solubility
Figure 8.4: Comparison of normalized methane solubility curves, c̃m,sol [z̃], com-
puted from rigorous thermodynamic models versus those fitted with equation
(8.51). Two different seafloor depths are considered, with the corresponding fitting
parameters, r2 , listed in the inset. A geotherm of 0.04◦ C/m, seafloor temperature
of 3◦ C, and seawater salinity were used for the solubility curves.
Figure 8.5 shows the steady-state sulfate concentration profiles, obtained through
equation (8.18), for three different scaled SMT depths. The sum Pe1 + Pe2 corre-
sponding to each scaled SMT depth is also shown. Figure 8.6 shows plots of the
normalized depth below the seafloor for the same scaled SMT depths as shown
in Figure 8.5. The solubility curve corresponding to seafloor depth of 1000 mbsl
(r2 = 0.625) is used for the results in Figures 8.5 and 8.6. Due to co-consumption
157
of sulfate and methane at the SMT, shorter L̃s indicates higher methane flux from
below, thereby leading to a shallower top of the gas hydrate layer (Figure 8.6).
0.005
Normalized depth
0.010
0.015
0.020
~
L = 0.01, Pe1 + Pe2 = − 15.5
~s
0.025 L = 0.02, Pe + Pe = − 7.0
~s 1 2
L = 0.03, Pe + Pe = − 4.3
s 1 2
0.030
0 0.2 0.4 0.6 0.8 1
Normalized sulfate concentration
Gas hydrate saturation profiles as a function of scaled SMT depths show higher
saturations within the GHSZ with decreasing L̃s , again due to net increase in
methane flux (Figure 8.7). Increase in the thickness of the hydrate layer with de-
creasing L̃s is also evident from the saturation profiles. We further compare steady-
state gas hydrate saturation profiles obtained from simulation results (crosses) of
Bhatnagar et al. (2008), which reveal good agreement between the theory devel-
158
Sulfate
0.05
0.10
Normalized depth
0.15
Methane
0.20
0.30
~
L = 0.01, Pe +Pe = − 15.5
~s 1 2
L = 0.02, Pe +Pe = − 7.0
0.35 ~ s 1 2
Ls = 0.03, Pe1+Pe2 = − 4.3
0.40
0 0.2 0.4 0.6 0.8 1
Normalized concentration
Figure 8.6: Effect of different SMT depths on steady-state pore water concentration
profiles. The methane solubility curve corresponds to seafloor depth of 1000m,
seafloor temperature of 3◦ C, and geotherm of 0.04◦ C/m. Specifying the scaled
SMT depth, L̃s , uniquely constrains the sulfate and methane concentration profiles,
as well as the top of the gas hydrate layer (intersection of the methane concentra-
tion profile with the solubility curve). Lower values of L̃s imply faster depletion of
sulfate due to higher methane flux from below, causing shallower occurrence of the
top of the gas hydrate layer.
oped in this paper and the numerical formulation. The profiles in Figures 8.6 and
8.7 clearly highlight that each distinct value of L̃s results in a unique profile for
We previously related the gas hydrate flux (Pe1 hSh i) to L̃s through equation
(8.44). This relationship is depicted in Figure 8.8 for two different seafloor depths
of 1000 mbsl and 3000 mbsl (same solubility curves as in Figure 8.4). Both curves
show that a decrease in L̃s leads to higher values of Pe1 hSh i due to higher net
159
0
~L = 0.01, Pe +Pe = − 15.5
~s 1 2
L = 0.02, Pe +Pe = − 7.0
0.2 ~s 1 2
L = 0.03, Pe +Pe = − 4.3
s 1 2
Normalized depth
Top of gas
0.4 hydrate layer
0.6
0.8
1.0
0 5 10 15
Gas hydrate saturation (% of pore volume)
Figure 8.7: Effect of variable SMT depths (same as in Figure 8.6) on steady-
state gas hydrate saturation profiles. Shallower SMT depths, indicating higher net
methane flux from depth, result in higher gas hydrate saturations within the GHSZ.
Calculation of gas hydrate saturation profile requires specification of Pe1 , which
was set equal to 0.1 for all three cases. Numerical simulation results (crosses)
from the model of Bhatnagar et al. (2008) match well with the analytical saturation
profiles (curves). Methane solubility curve used is the same as in Figure 8.6.
methane input to the system. Both set of curves also truncate at a maximum
scaled SMT depth (L̃s,max ), beyond which the methane flux is too low to form any
gas hydrate. L̃s,max for the solubility curves corresponding to seafloor depths of
1000 mbsl and 3000 mbsl are equal to 0.059 and 0.058, respectively. Again, results
from our analytical model match well with simulations performed for the same set
1.5
Analytical solution (Seafloor depth 1000 mbsl)
Analytical solution (Seafloor depth 3000 mbsl)
Numerical solution (Seafloor depth 1000 mbsl)
1.0
0.5
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
~
Scaled SMT depth: Ls
Figure 8.8: Relationship between average gas hydrate flux (Pe1 hSh i) and scaled
SMT depth (L̃s ) for different seafloor depths. Curves representing analytical solu-
tions are also compared with steady state numerical simulations of Bhatnagar et al.
(2008). Shallow SMT depths indicate higher methane flux from deeper sources
causing higher average gas hydrate flux (and saturations) through the GHSZ.
Several gas hydrate settings show correlation between the SMT depth and the
depth to the shallowest occurrence of gas hydrate (e.g., Borowski et al., 1999),
with some sites suggesting gas hydrate first occurring at a depth 10 times the
SMT depth (e.g., Paull et al., 2005). Our model allows the estimation of this ratio
[(Lt − Lh )/Ls ], or [(1 − L̃h )/L̃s ] in scaled form, as a function of the system param-
eters. Figure 8.9 shows this ratio as a function of L̃s for the two solubility curves
shown in Figure 8.4. Overall, the ratio increases with decreasing L̃s , though the
161
variation with scaled SMT depth is more gradual at relatively large values of L̃s .
However, this ratio increases rapidly with decreasing L̃s at relatively small values
of L̃s , implying that with increasing methane flux from below, the SMT migrates
upwards faster than the increase in thickness of the gas hydrate layer (Lh ). Except
at gas vents or sites neighboring rapid fluid flow conduits, the scaled SMT depth at
most gas hydrate settings is usually greater than about 0.01, which indicates that
the top of the gas hydrate layer occurring at about 10−12 times the SMT depth
30
Solubility curve for seafloor depth 1000 mbsl
Solubility curve for seafloor depth 3000 mbsl
25
s
20
Ratio: (L − L )/L
h
15
t
10
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
Scaled SMT depth: Ls/Lt
Figure 8.9: Relationship between the ratio of depth to the first occurrence of gas
hydrate to the SMT depth as a function of the scaled SMT depth for two differ-
ent solubility curves (Figure 8.4). This ratio is close to 10−12 for relatively large
SMT depths but significantly departs from this suggested range for shallower SMT
depths. Similar to Figure 8.8, both curves truncate at the maximum allowed SMT
depth for each case.
162
ward fluid flow and localized areas of gas venting (Tréhu et al., 2004; Riedel et al.,
2006). Results from Ocean Drilling Program Leg 204 and Integrated Ocean Drilling
Program Expedition 311 have given great insight into the complex and heteroge-
neous gas hydrate distribution at several sites drilled along this margin (Tréhu et al.,
2003, 2004; Riedel et al., 2006). Sites in this region are characterized by relatively
high fluid fluxes and low average total organic carbon (TOC) content (Westbrook
et al., 1994; Riedel et al., 2006), implying that fluids from depth form the dominant
methane source. This makes sites along Cascadia Margin a good location to test
our model. We use SMT depths and other data for four Cascadia Margin sites
(Table 8.1) to predict gas hydrate saturations, average saturation and depth to the
first occurrence of gas hydrate below the seafloor (Table 8.2). These sites include
Site 889, drilled as part of ODP Leg 146, and Sites U1325, U1326 and U1329 that
Table 8.1 lists the site number, sedimentation rate (Ṡ), seafloor temperature
(T ), geothermal gradient (G), seafloor depth (D0 ), dimensional SMT depth (Ls ),
dimensional BSR depth (Lt ), methane solubility in water at the base of the GHSZ
(clm,eqb ), constant m (equation (8.22)), and the fitting constant to the solubility curve
r2 (equation (8.51)) for all four sites. For each site, the methane solubility curve
is generated for the corresponding seafloor conditions using the rigorous thermo-
dynamic model of Bhatnagar et al. (2007). These calculations also give clm,eqb for
163
Site Ṡ T0 G D0 Ls Lt clm,eqb m r2
(cm/k.y.) (o C) (o C/m) (m) (m) (m)
U1329 9.2 3.3 0.072 946 9.4 126 1.8×10−3 3.8 0.52
each site, from which the constant m can be computed. The generated solubility
explain how various columns in Table 8.2 are sequentially obtained from site-
specific data. Using the sedimentation rate and porosity, the first Peclet number
(Pe1 ) is computed for each site. Ratio of the dimensional depths Ls and Lt give
the scaled SMT depth L̃s . Using this value, and other site-specific parameters, the
maximum SMT depth for hydrates to form (L̃s,max ) is computed. Solution of equa-
tions (8.38) and (8.39) gives methane flux and thickness of the gas hydrate layer
from which the scaled depth to the first occurrence of gas hydrate (1 − L̃h ) is calcu-
lated. Multiplying this by Lt yields the dimensional depth to the first occurrence of
hydrate (Lt − Lh ). Using Pe1 characteristic of each site, we next compute the gas
hydrate saturation profile (equation (8.41)). Averages of this gas hydrate saturation
164
Site Pe1 L̃s L̃s,max Lt − Lh hSh iGHSZ hSh iGHSZ hSh iGHSZ
(m) (calc.) (res. log) (Cl− )
profile over the entire GHSZ as well as the gas hydrate occurrence zone (GHOZ)
are also computed. Finally, average gas hydrate saturations at these sites from
other proxy data (resistivity log and chloride) are listed. Figure 8.10 shows the gas
hydrate saturation profiles at three of these sites. Site U1329 is not shown on this
Site 889 (ODP Leg 146) has been previously modeled as a gas hydrate system
dominated by deeper methane sources (Davie and Buffett, 2003a; Bhatnagar et al.,
2007). Davie and Buffett (2003a) fit the pore water chloride profile at Site 889 using
a coupled numerical model with methane supply from depth. Their results indicate
peak hydrate saturation close to 2% at the base of GHSZ and average saturation
<1% within the GHSZ (Davie and Buffett, 2003a). This result agrees favorably
with our simulation that shows peak saturation of about 2.7% at the base of GHSZ
(Figure 8.10) and average saturation of 0.6% across the entire GHSZ (Table 8.2).
165
Hyndman et al. (1999) calculated gas hydrate saturation between 25-30% of pore
space in the 100 m interval above the base of GHSZ at Site 889 using resistivity
log data. However, subsequent calculations using a different set of Archie param-
eters have revised this estimate to 5-10% in that 100 m interval (Collett, 2000).
Further, Ussler and Paull (2001) show that a smoothly decreasing chlorinity profile
at Site 889 yields gas hydrate saturation of 2-5% within discrete layers. Temper-
889 (Kastner et al., 1995). Although several parameter uncertainties confront such
geochemical and geophysical estimates (Egeberg and Dickens, 1999; Riedel et al.,
2006), average saturation predicted using our SMT based model concurs with the
For the IODP Expedition 311 sites, drilled along the northern Cascadia Mar-
gin, we compare our predictions with average saturations computed from chloride
anomalies and resistivity log data (Table 8.2). Average saturation is calculated from
chloride data by assuming a background in situ chloride profile and attributing the
relative pore water freshening to gas hydrate dissociation (e.g., (Egeberg and Dick-
ens, 1999)). Average saturation is obtained from resistivity data using the Archie
equation and parameters given in Riedel et al. (2006). Average saturation over
the GHSZ at Site U1325 is estimated from resistivity data and chloride anomalies
chlorinity for Site U1326 are 6.7% and 5.5%, respectively. These values compare
favorably with 3.1% and 6.6% average saturation from our SMT based model at
Figure 8.10: Steady-state gas hydrate saturation profiles computed from scaled
SMT depths at Cascadia Margin Sites 889, U1325 and U1326. Scaled SMT depth
is highest for Site 889 and lowest for Site U1326, implying higher methane flux and
greater gas hydrate saturation at Site U1326 and relatively low methane flux and
hydrate saturation at Site 889.
At Site U1329, our model predicts that the SMT is greater than the maximum
allowed, implying that methane flux is too low at this site to precipitate any gas
hydrate. Although resistivity data suggests a small amount of gas hydrate within
the GHSZ, our prediction is supported by the very low saturation predicted from
chlorinity data. Site U1329 is supposed to mark the eastern limit of gas hydrate
occurrence along this transect and has a very faint BSR (Riedel et al., 2006). Fur-
ther, no gas hydrate was recovered from cores at this site. These features indicate
that, in general, Site U1329 has no hydrate or very little hydrate within some local-
167
ized layers.
In general, we get good first order agreement between average saturations de-
rived using resistivity logs/chloride anomalies and those predicted using our model,
although our model consistently predicts lower average saturation at all four sites
along Cascadia Margin. For estimates from resistivity logs, a possible explanation
formation water resistivity and three empirical constants, which are hard to con-
uration at Site 889 from resistivity data have since then been revised to much lower
values using new Archie parameters (Riedel et al., 2006). Moreover, most studies
employing transport models (e.g., (Xu and Ruppel, 1999; Davie and Buffett, 2003a;
Bhatnagar et al., 2007)) predict gas hydrate to first occur well below the seafloor. In
contrast, log-based results often predict gas hydrate starting immediately below the
seafloor. This will cause saturations from transport models to be lower than those
predicted using resistivity log data. Similarly, estimation of hydrate saturation from
chloride data is quite sensitive to the choice of baseline curves. Apart from the
small deviations between model and chloride/resistivity log predictions, our model
gives a good average estimate of gas hydrate saturation. Further, it captures the
variation in average saturation across the IODP Expedition 311 transect correctly
8.6 Caveats
sources. Compared to full numerical simulations, this method gives a quick and
• The above analysis only applies to gas hydrate systems where methane from
depth is the only source. Gas hydrate settings with significant organic carbon
input can deplete pore water sulfate by bacteria that utilize organic carbon
instead of methane (Berner, 1980). Thus, if some of the pore water sulfate is
depleted through this alternate pathway, the resulting SMT will become shal-
lower than the SMT due to methane flux alone. In such cases, our method
will yield gas hydrate saturations higher than the actual saturation.
• Methane is the only hydrate forming gas in our model. In some regions (e.g.,
Gulf of Mexico), upward fluid flux might contain higher alkanes that can also
consume sulfate (Joye et al., 2004), making the 1:1 sulfate-methane flux re-
sites dominated by focused fluid flow, possibly along high permeability con-
duits, will probably show greater deviations between gas hydrate saturation
169
estimated from our model and that inferred from other proxy data (e.g., re-
dict gas hydrate distributed heterogeneously within the GHSZ that is possibly
controlled by the local lithology. Our simple formulation does not include such
age sense.
• Our model does not account for free gas existence or migration within the
GHSZ. Several sites associated with free gas venting have been drilled along
Cascadia Margin (e.g., Sites 1249 and 1250, ODP Leg 204). These sites
show massive layers of hydrate close to the seafloor. Due to relatively high
methane flux, such sites might not have any detectable pore water sulfate
(Tréhu et al., 2003). Further, gas hydrate saturation at such sites decreases
with depth below the seafloor (Haeckel et al., 2004; Liu and Flemings, 2006),
gas hydrate saturation with depth below the seafloor (Xu and Ruppel, 1999;
Davie and Buffett, 2001; Bhatnagar et al., 2007, 2008). Predictions from our
analytical model agree with this latter category of models. Consequently, the
formulation developed here will not predict massive gas hydrate layers close
8.7 Conclusions
saturation from scaled depth of the sulfate-methane transition (SMT) for gas hy-
the ratio of the dimensional depth of the SMT below the seafloor to the depth
of the gas hydrate stability zone (GHSZ) below the seafloor. Using simple one-
dimensional mass balances for sulfate and methane, we show that net methane
flux in such deep-source systems uniquely determines the scaled SMT depth, the
thickness of the gas hydrate layer and steady-state gas hydrate saturation within
the GHSZ. This dependence allows estimation of the vertical extent of gas hydrate
and its saturation through knowledge of SMT depth at a given site. Steady-state
results show that as the SMT becomes shallower, methane flux and, consequently,
Our formulation also gives an estimate of the maximum scaled SMT depth and
minimum gas hydrate thickness (or, equivalently, minimum methane flux) for gas
hydrates to be present at any given setting. Any scaled SMT depth greater than
this maximum implies that methane flux is lower than the minimum required to form
hydrate. The one-to-one relationship between SMT depth and thickness of the gas
hydrate layer allows us to obtain the ratio of the first occurrence of gas hydrate
below the seafloor to the SMT depth. We show that this ratio can be about 10-
higher methane fluxes will cause this ratio to increase towards much higher values.
171
Average saturations over the GHSZ at four Cascadia Margin locations, calcu-
lated from our method, are 0.6%, 2.3%, 5.5% and 0% for Sites 889, U1325, U1326
and U1329, respectively. These values compare favorably with averages computed
from resistivity log and chlorinity data for all sites. Hence, our analytical formula-
tion provides a simple and fast technique to constrain gas hydrate saturation in
Chapter 9
9.1 Introduction
Marine gas hydrate systems are often characterized by overpressure, i.e. pore wa-
ter pressures higher than hydrostatic. This is particularly evident at settings dom-
inated by low permeability silts/clays, e.g. Blake Ridge (Paull et al., 2000; Flem-
1958). Overpressure can develop in such systems if pore water cannot be ex-
pelled from the pore space fast enough and, instead, starts to support some of the
overburden stress. Since permeability controls this rate of pore water expulsion,
sediments with low permeability can develop overpressure (Wangen, 1992; Dugan
with relatively high permeability if the sedimentation rate is fast, i.e. increase in
its obvious effect on fluid flow (Dugan and Flemings, 2000; Flemings et al., 2003;
Gering, 2003). However, overpressure also impacts the behavior of such systems
in less intuitive ways. For example, the thickness of the free gas layer below the
173
base of the gas hydrate stability zone (GHSZ) depends on the magnitude of over-
pressure. The free gas saturation profile associated with marine gas hydrate sys-
phase (Davie and Buffett, 2001, 2003a; Bhatnagar et al., 2007). However, when
free gas saturation exceeds its critical saturation, it no longer moves as a trapped
phase within the sediment matrix. Instead, free gas forms a connected column that
is sealed by the high capillary entry pressure at the base of the GHSZ due to gas
hydrate (Liu and Flemings, 2007). The pressure within this connected gas column
increases towards the base of the GHSZ, whereas the length of this gas column
is regulated by the difference between pore water pressure and lithostatic stress.
Once gas pressures reach this lithostatic limit at the base of the GHSZ, the vertical
gas effective stress becomes zero. Figure 9.1 illustrates this mechanism schemat-
relatively low, thereby allowing a large gas column to form before maximum gas
pressure at the top of the connected column equals the lithostatic stress (overbur-
den). When gas pressure exceeds the lithostatic limit, fractures can dilate, allowing
free gas migration into the GHSZ (Flemings et al., 2003; Hornbach et al., 2003;
Tréhu et al., 2004). Sediments might, however, fail at much lower gas pressures
due to shear failure (Finkbeiner et al., 2001; Hornbach et al., 2003). Compared
to this scenario, Figure 9.1b shows a case where pore pressure gradient deviates
significantly from the hydrostatic gradient. This results in higher pore water pres-
sure beneath the GHSZ, thereby reducing the maximum thickness of the free gas
column. Thus, the thickness of the connected free gas layer is regulated by the
174
Plith Plith
Depth
Depth
BHSZ BHSZ
P
w
P
g
Phydro P
hydro
Pg
P
w
Pressure Pressure
dynamics.
and gas hydrate velocity through the GHSZ. At hydrostatic pore pressures, sedi-
ments achieve maximum compaction due to high effective stresses acting on them.
ties, which leads to faster sediment velocities through the GHSZ. This curtails the
amount of organic carbon converted within the GHSZ as well as the residence time
of gas hydrate in this zone, resulting in lower gas hydrate and free gas saturations.
and extends the base of the GHSZ to greater depths below the seafloor. This will
possibly lead to deeper gas hydrate occurrences below the seafloor. This might
GHSZ, assuming hydrostatic pressure, and that obtained from a seismic BSR.
However, we show through numerical simulations that this increase in depth of the
GHSZ due to overpressure is relatively small, even when pore pressures become
predicted depths of GHSZ may not be reliable indicators of in situ pore pressures
generation in marine gas hydrate systems and study the parameters governing
make use of Darcy’s law to model fluid flow relative to the compacting sediment.
As mentioned before, free gas becomes mobile only when its saturation exceeds
some critical value, which can be about 5-10% gas saturation. We focus on the
meability in this chapter and consequently assume a relatively higher value of crit-
ical gas saturation of 10%. This ensures that free gas will remain immobile within
the sediment matrix. Mobile free gas will be modeled in the next chapter.
The water, methane, sediment and organic mass balances are written as:
Water Balance
∂ h i ∂ h i
φSw clw ρw + φSh chw ρh + φSw clw ρw vw + φSh chw ρh vs = 0 (9.1)
∂t ∂z
Methane Balance
∂ h i
φSw clm ρw + φSh chm ρh + φSg cgm ρg +
∂t
177
" #
∂ h i ∂ ∂cl
φSw clm ρw vw + φSh chm ρh vs + φSg cgm ρg vg = φSw Dm ρw m
∂z ∂z ∂z
MCH4
+ ρs λ (1 − φ) α (9.2)
Morg
Sediment Balance
∂ ∂
[(1 − φ) ρs ] + [(1 − φ) ρs vs ] = 0 (9.3)
∂t ∂z
Organic Balance
∂ ∂
[(1 − φ) ρs α] + [(1 − φ) ρs vs α] = −ρs λ (1 − φ) α (9.4)
∂t ∂z
iments.
à !
kkrw ∂pw
Sw φ(vw − vs ) = − − ρw g (9.5)
µw ∂z
water.
where k0 and φ0 are the initial sediment permeability and porosity at time of depo-
sition.
• Water relative permeability model in the presence of gas hydrate is (with pore-
2(1 − Sh )2
krw = 1 − Sh2 + (9.7)
ln(Sh )
• Water relative permeability model in the presence of free gas is (Bear, 1988):
0 Sw − Swr
krw = krw (Sw∗ )4 , where Sw∗ = (9.8)
1 − Swr
− σvσ−pw
φ = φ∞ + (φ0 − φ∞ )e φ (9.9)
∂σv
= [(1 − φ)ρs + φρw ] g (9.10)
∂z
179
The above equations are now written in dimensionless form. Reduced porosities
φ − φ∞ φ0 − φ∞ 1 − φ∞
φ̃ = , η= , γ= (9.11)
1 − φ∞ 1 − φ∞ φ∞
The dimensionless Peclet number and Damkohler number are defined as be-
fore:
Uf,sed Lt λL2t
Pe1 = , Da = (9.12)
Dm Dm
and sedimentation rate, thus quantifying the effects of sedimentation and com-
Larger values of Nsc correspond to higher absolute permeability and/or low sedi-
mentation rate, implying hydrostatic pressures. Conversely, low Nsc values imply
low sediment permeability and/or high sedimentation rate, thereby causing pore
pressures higher than hydrostatic. Dimensionless groups similar to Nsc have been
group Ntφ :
Lφ σφ /(ρw g)
Ntφ = = (9.14)
Lt Lt
Lithostatic stress (σv ), water pressure and gas pressure are normalized by hydro-
σv pw pg
σ̃v = , p̃w = , p̃g = (9.16)
ρw gLt ρw gLt ρw gLt
z t
z̃ = , t̃ = (9.17)
Lt L2t /Dm
ρh ρg ρs
ρ̃h = , ρ̃g = , ρ̃s = (9.18)
ρw ρw ρw
Ṡ:
vs
ṽs = (9.19)
Ṡ
181
Finally, organic carbon content and initial TOC are scaled as:
α α0
α̃ = , β= l
(9.20)
α0 cm,eqb
The above scaling schemes lead to the following dimensionless form of the four
Water Balance:
" #
∂ 1 + γ φ̃ ³ ´
(1 − Sh − Sg )(1 − clm,eqb c̃lm ) + Sh chw ρ̃h +
∂ t̃ γ
à ! ·
1+γ ∂ 1 + γ φ̃
Pe1 (1 − Sh − Sg )(1 − clm,eqb c̃lm )ṽs −
1−η ∂ z̃ γ
à !8 à ! ¸
1+γ 1 + γ φ̃ ∂ p̃w 1 + γ φ̃
Nsc (1 − clm,eqb c̃lm )krw −1 + Sh chw ρ̃h ṽs = 0 (9.21)
γ 1 + γη ∂ z̃ γ
ρw gD0 + ρw gz D0
I.C. : p̃w (z̃, 0) = = + z̃ (9.22)
ρw gLt Lt
D0
B.C.(1) : p̃w (0, t̃) = (9.23)
Lt
∂ p̃w
B.C.(2) : (D, t̃) = 1 , (Hydrostatic gradient) (9.24)
∂ z̃
Methane balance:
" #
∂ 1 + γ φ̃ ³ ´
(1 − Sh − Sg )c̃lm + Sh c̃hm ρ̃h + Sg c̃gm ρ̃g +
∂ t̃ γ
182
à ! "
1+γ ∂ 1 + γ φ̃
Pe1 (1 − Sh − Sg )c̃lm ṽs +
1−η ∂ z̃ γ
à ! à !8 à !
1 + γ φ̃ 1+γ 1 + γ φ̃ ∂ p̃w
Sh c̃hm ρ̃h ṽs − Nsc krw c̃lm −1 +
γ γ 1 + γη ∂ z̃
à ! à !8 à !à !#
1 + γ φ̃ 1+γ 1 + γ φ̃ µw ∂ p̃g
Sg c̃gm ρ̃g ṽs − Nsc krg ρ̃g c̃gm − ρ̃g =
γ γ 1 + γη µg ∂ z̃
" #
∂ 1 + γ φ̃ ∂c̃l MCH4
(1 − Sh − Sg ) m + ρ̃s Da(1 − φ̃)α̃β (9.25)
∂ z̃ γ ∂ z̃ Morg
∂c̃lm
B.C.(2) : (D, t̃) = 0 (9.28)
∂ z̃
Sediment balance:
à !
∂ h i 1+γ ∂ h i
1 − φ̃ + Pe1 (1 − φ̃)ṽs = 0 (9.29)
∂ t̃ 1−η ∂ z̃
1−η
I.C. : ṽs (z̃, 0) = , (for hydrostatic compaction) (9.30)
1 − φ̃
Organic balance:
à !
∂ h³ ´ i 1+γ ∂ h i
1 − φ̃ α̃ + Pe1 (1 − φ̃)ṽs α̃
∂ t̃ 1−η ∂ z̃
³ ´
= −Da 1 − φ̃ α̃ (9.32)
Reduced porosity φ̃ is related to the dimensionless lithostatic stress (σ̃) and dimen-
" #
σ̃v − p̃w
φ̃ = η exp − (9.35)
Ntφ
η
φ̃ = ³
γ z̃(ρ̃s −1)
´ (9.36)
η + (1 − η) exp Ntφ (1+γ)
∂ σ̃v 1 h ³ ´ ³ ´i
= ρ̃s γ 1 − φ̃ + 1 + γ φ̃ (9.37)
∂ z̃ 1+γ
The coupled equations (9.21), (9.25), (9.29) and (9.32) are solved numerically us-
ing a fully implicit formulation, with the primary variables being p̃w , ṽs , α̃ and one
of the following three (c̃lm ,Sh , Sg ). Choice between these last three primary vari-
at any given timestep. All four mass balances are cast in residual form and the
9.3 Results
To study the effect of overpressure on gas hydrate and free gas saturation, we
Nsc . Apart from the parameter Nsc , other primary simulation parameters include
the Peclet number, the Damkohler number, the normalized organic carbon input
and the reduced porosity parameters. Values of these parameters used in all sim-
ulations shown in this chapter are: Pe1 = 0.1, Da = 10, β = 3, Ntφ = 1, η = 6/9,
and γ = 9. Seafloor parameters are chosen to be similar to the Blake Ridge region
(Paull et al., 2000), with seafloor temperature of 3◦ C, seafloor depth of 2700 m, and
later and only vary Nsc from higher to progressively lower values.
Figure 9.2 shows steady-state pore pressure profiles versus depth below the
seafloor for four different values of Nsc . Hydrostatic pressure profile and lithostatic
stress profiles are also plotted for each case as minimum and maximum bounds to
the pore pressure, respectively. Importantly, simulations show that relatively higher
values of Nsc (∼ 104 ) lead to almost hydrostatic pore pressures, whereas relatively
lower Nsc (of order unity) lead to pore pressures that are close to the lithostatic limit
(Figure 9.2). This occurs because relatively low Nsc values imply lower sediment
permeability and/or fast sedimentation rate. Either of these conditions can reduce
185
0 0
Nsc = 10000 N = 100
Normalized depth
Normalized depth
sc
0.5 0.5
P
lith
1.0 1.0
Plith
Phydro P
1.5 Pw 1.5 hydro Pw
2.0 2.0
5 6 7 8 9 10 5 6 7 8 9 10
Normalized pressure Normalized pressure
0 0
Nsc = 10
Normalized depth
Nsc = 2
Normalized depth
0.5 0.5
Plith
1.0 1.0
Phydro
Phydro P
1.5 1.5 lith
P Pw
w
2.0 2.0
5 6 7 8 9 10 5 6 7 8 9 10
Normalized pressure Normalized pressure
pore pressures higher than hydrostatic values. Conversely, relatively higher values
of Nsc imply high sediment permeability and/or low sedimentation rate. This main-
tains equilibrium compaction and pore pressures that remain close to hydrostatic
values.
0
N = 10000
sc
0.2 N = 100
sc
Nsc = 10
0.4
N =2
sc
Normalized depth
0.6
0.8
1.0
1.2
1.4
Hydrostatic
1.6 profile
1.8
2.0
5 6 7 8 9 10
Normalized pressure
Figure 9.3: Pressure profiles shown in Figure 9.2 plotted together for all four cases.
Lithostatic stress profiles (dashed curves) and the corresponding pore pressure
profiles (solid curves) for the same value of Nsc are color-coded together.
Figure 9.2 also reveals that the lithostatic stress reduces as pore pressures
increase. To illustrate this more clearly, we plot pore pressure and lithostatic stress
profiles corresponding to the four cases together in Figure 9.3. Dashed set of
curves in Figure 9.3 correspond to the lithostatic stress profiles, while solid curves
187
denote pore pressure profiles. It can be seen from Figure 9.3 that pore pressure
and lithostatic profiles remain separated from each other at large Nsc . However,
on lowering Nsc , both pore pressure and lithostatic curves approach each other.
resulting from lower effective stresses acting on the sediments, cause lower bulk
sediment bulk density, overpressure and higher porosities lead to lower lithostatic
stress (Figure 9.3). Steady-state reduced porosity profiles for the four cases are
plotted in Figure 9.4. As expected, higher pore pressures at low Nsc lead to lower
stability of gas hydrates. Specifically, increase in pore pressure extends the depth
to the base of the GHSZ deeper into the sediments. This change is shown through
the methane solubility curves in Figure 9.5 for the same set of Nsc values simulated
in the previous figures. We start with the case Nsc =10000, which corresponds to
ously, the solubility curve for this case has a peak methane solubility equal to unity
at unit normalized depth. As pore pressure increases, i.e. Nsc decreases, we ob-
serve that the peak values of the solubility curves shift to higher values, with the
peak itself occurring at slightly deeper depths. This demonstrates that the base
However, even when pore pressures are close to lithostatic, the downward shift in
the base of the GHSZ is very small in the normalized form. When this normalized
188
0
N =2
sc
0.2 N = 10
sc
Nsc = 100
0.4
N = 10000
sc
Normalized depth 0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0 0.2 0.4 0.6 0.8 1
~
Reduced porosity φ
Figure 9.4: Effect of the parameter Nsc on steady-state reduced porosity profiles.
At relatively high values of Nsc , pore pressure remains close to hydrostatic, caus-
ing equilibrium compaction and maximum porosity loss. As Nsc is reduced, pore
pressure increases from hydrostatic values, resulting in lower effective stresses,
lesser compaction and significantly higher porosities.
depth scale is converted back to the physical scale by multiplying with Lt , the depth
to the base of GHSZ, this increase in the thickness of the GHSZ becomes larger,
but is still relatively small. For example, for the case corresponding to almost litho-
static pore pressure, the downward shift in the base of the GHSZ is about 20 m,
which, for Blake Ridge type seafloor conditions, is only about 0.7% of the water
depth.
The effect of Nsc on steady-state gas hydrate and free gas saturation profiles is
shown in Figure 9.6. Maximum gas hydrate and free gas saturations occur at the
0
N = 10000
sc
0.2 N = 100
sc
Nsc = 10
0.4
N =2
sc
0.6
Normalized depth
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0 0.2 0.4 0.6 0.8 1 1.2
Normalized methane solubility
decrease in the values of Nsc lead to lower gas hydrate and free gas saturations
Nsc lead to higher overpressures, higher sediment porosities and faster sediment
velocities, which result in lower organic carbon decay within the GHSZ and shorter
However, it should be noted that hydrate and free gas saturation profiles do
not give a complete picture of their amounts, because each value of Nsc results in
a different porosity profile. Thus, although hydrate and free gas saturation within
the pore space decrease on lowering Nsc , the corresponding increase in porosity
190
0
N = 10000
sc
0.2 N = 100
sc
Nsc = 10
0.4
N =2
sc
Normalized depth
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Phase Saturation
might lead to net higher accumulation of hydrate or free gas within the sediment
volume. To test this scenario, we plot the product of porosity and hydrate/free gas
saturation (φSj ) to get the volume fraction of hydrate and free gas within the sed-
iment. These profiles are plotted in Figure 9.7 and show that the net amount of
gas hydrate or free gas saturation within the sediment also decreases on lowering
Nsc . However, multiplying by porosity does reduce the magnitude of change ob-
served between different cases. For example, peak hydrate saturation at the base
0
N = 10000
sc
0.2 N = 100
sc
Nsc = 10
0.4
N =2
sc
0.6
Normalized depth
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0 0.005 0.01 0.015 0.02 0.025
Sediment volume fraction (φ S )
j
goes from about 2.2% to 0.7%, a factor of 3 change, for the same decrease in Nsc .
This validates our hypothesis that overpressure does lower the net amount of
methane that accumulates in either hydrate or free gas phase. In other words,
the decrease in hydrate and free gas saturation with increasing overpressure is
not only a result of increased porosities. The effect of increased sediment and
fluid velocities and lower organic carbon decay within the GHSZ has a much more
Based on pressure profiles shown in Figures 9.1 and 9.2, it can be argued that
deep connected free gas columns may result for settings characterized by high
Nsc . In contrast, only short gas columns can form when Nsc is low before sedi-
ment fracture/failure occurs and vents the free gas into the ocean. Thus, from an
sedimentation rates (i.e., high Nsc ) might be most suitable for targeting the free gas
9.4 Conclusions
veloped to study the effect of overpressure on gas hydrate and free gas satura-
Simulations show that relatively high values of Nsc (close to 104 ) lead to systems
close to hydrostatic pore pressure, while relatively low values of Nsc (about unity)
this gas hydrate system by lowering effective stresses on the sediment, causing
high porosities ultimately lead to lesser organic carbon decay, resulting in lower hy-
drate and free gas saturations for our set of boundary conditions, i.e., fixed seafloor
Chapter 10
10.1 Introduction
tion of key dimensionless groups controlling the system and helped in summarizing
propriate scaling schemes (Bhatnagar et al., 2007). However, natural gas hydrate
systems are much more complex and heterogeneous than these simplified 1-D
• Free gas migration: Free gas is assumed to be immobile in most gas hydrate
models. However, gas hydrate can become mobile if it exceeds the critical
ment of thick gas columns can result in sediment failure and venting of free
gas into the gas hydrate stability zone (GHSZ) or the ocean (Flemings et al.,
erogeneities. This can help predict high saturation hydrate/free gas accumu-
• Faulted or fractured systems: Fluid flow within natural gas hydrate systems
ity conduits can preferentially feed methane-charged water and gas to high
iment layer. Geochemical and geophysical proxy data often show relatively
high gas hydrate saturation within local sand beds that alternate with silt or
clay layers with little or no hydrate (Paull et al., 2000; Tréhu et al., 2003; Riedel
et al., 2006; Weinberger and Brown, 2006). Effects of such parallel or dipping
To incorporate these features, we extend the previous 1-D model to 2-D sys-
tems with mobile free gas in this chapter. Heterogeneities, such as vertical frac-
tures and parallel or dipping sediment layers are then added to this 2-D model to
simulate complex and real-world gas hydrate settings. We start with component
The two-dimensional material balances for water, sediment, organic carbon and
Water Balance
∂ h i h i
φSw clw ρw + φSh chw ρh + ∇ · φSw clw ρw vw + φSh chw ρh vs = 0 (10.1)
∂t
Sediment Balance
∂
[(1 − φ) ρs ] + ∇ · [(1 − φ) ρs vs ] = 0 (10.2)
∂t
Organic Balance
∂
[(1 − φ) ρs α] + ∇ · [(1 − φ) ρs vs α] = −ρs λ (1 − φ) α (10.3)
∂t
Methane Balance
∂ h i
φSw clm ρw + φSh chm ρh + φSg cgm ρg +
∂t
h i h i
∇ · φSw clm ρw vw + φSh chm ρh vs + φSg cgm ρg vg = ∇ · φSw Dm ρw ∇clm
MCH4
+ ρs λ (1 − φ) α (10.4)
Morg
g denote methane, water, sediment, hydrate, and free gas, respectively. Super-
scripts l, h and g represent water, hydrate and free gas phases, respectively.
kkrw
Sw φ(vw − vs ) = − (∇pw − ρw g∇z) (10.5)
µw
where k is the absolute permeability of the sediment, krw denotes relative perme-
kkrg
Sg φ(vg − vs ) = − (∇pg − ρg g∇z) (10.6)
µg
where krg denotes relative permeability of gas, µg is gas viscosity, and pg is gas
pressure.
where k0 and φ0 are the initial sediment permeability and porosity at time of depo-
197
sition.
à !
2(1 − Sh )2
k = k0 1 − Sh2 + (10.8)
ln(Sh )
This function (k/k0 ) is plotted in Figure 10.1 as a function of gas hydrate saturation,
Sh . The ratio approaches unity at hydrate saturation close to zero and decreases
increases the capillary entry pressure for free gas, effectively making gas hydrate
• Water relative permeability model in the presence of free gas is (Bear, 1988):
0 Sw − Swr
krw = krw (Sw∗ )4 , where Sw∗ = (10.9)
1 − Swr
0
where krw is the end-point relative permeability and Swr is the residual water satu-
ration.
0 Sg − Sgr
krg = krg (Sg∗ )2 , where Sg∗ = (10.10)
1 − Swr − Sgr
0
where krg is the end-point relative permeability and Sgr is the residual gas satura-
tion.
198
0.9
0.8
0.7
0.6
k/k0
0.5
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
Gas hydrate saturation S
h
Figure 10.1: Decrease in sediment absolute permeability (k/k0 ) due to gas hy-
drate precipitation and assuming it to be pore-filling. Permeability to flow rapidly
decreases as hydrate saturation increases from zero.
• Capillary pressure (Pc ) is the difference between gas and water pressure:
Pc = pg − pw (10.11)
s
Pc,0 (Sw ) k0
J(Sw ) = (10.12)
σgw cos θ φ0
where Pc,0 (Sw ) is the capillary pressure curve at reference porosity φ0 and absolute
199
permeability k0 , σgw is the gas-water interfacial tension and θ is the contact angle.
Assuming constant σgw and θ, capillary pressure curve for any given porosity (φ)
s s
φ k0 φ
Pc (Sw ) = J(Sw )σgw cos θ = Pc,0 (Sw ) (10.13)
k φ0 k
The Brooks-Corey model is used to define the capillary pressure curve as a func-
Sw − Swr
Pc,0 (Sw ) = Pce,0 Sw∗ −1/n , where Sw∗ = (10.14)
1 − Swr
where Pce,0 is the capillary entry pressure at φ0 and k0 , and n is the pore-size
distribution index.
− σvσ−pw
φ = φ∞ + (φ0 − φ∞ )e φ (10.15)
∂σv
= [(1 − φ)ρs + φρw ] g (10.16)
∂z
200
The above equations are now written in dimensionless form according to the scal-
ing schemes defined in section 9.2.3, but are repeated here for continuity. Reduced
φ − φ∞ φ0 − φ∞ 1 − φ∞
φ̃ = , η= , γ= (10.17)
1 − φ∞ 1 − φ∞ φ∞
The dimensionless Peclet number, Damkohler number and the group quantify-
Uf,sed Lt λL2t k0 ρ w g
Pe1 = , Da = , Nsc = (10.18)
Dm Dm µw Ṡ
pressure, Lt is depth to the base of the GHSZ, and Ṡ is the sedimentation rate at
Ntφ :
Lφ σφ /(ρw g)
Ntφ = = (10.19)
Lt Lt
Lithostatic stress (σv ), water pressure, gas pressure and capillary pressure are
201
σv pw pg Pc
σ̃v = , p̃w = , p̃g = , P̃c = (10.21)
ρw gLt ρw gLt ρw gLt ρw gLt
z x t
z̃ = , x̃ = , t̃ = (10.22)
Lt Lt L2t /Dm
ρh ρg ρs
ρ̃h = , ρ̃g = , ρ̃s = (10.23)
ρw ρw ρw
zero lateral strain. This causes the sediment to move only in the vertical direction,
with no lateral movement. Consequently, gas hydrate within the matrix as well as
immobile free gas (if it is below critical saturation) can only move downward with
the sediment. Water and mobile gas, however, can move in vertical as well as
vs
ṽs = (10.24)
Ṡ
202
Finally, organic carbon content and initial TOC are scaled as:
α α0
α̃ = , β= l
(10.25)
α0 cm,eqb
where α0 is TOC content at the seafloor and clm,eqb is methane solubility in water at
The above scaling schemes lead to the following dimensionless form of the four
Water Balance:
" # Ã ! "(Ã !
∂ 1 + γ φ̃ ³ ´ 1+γ ∂ 1 + γ φ̃
Sw clw + Sh chw ρ̃h + Pe1 Sw ṽs −
∂ t̃ γ 1−η ∂ z̃ γ
à !à !8 à !) à ! #
1+γ 1 + γ φ̃ ∂ p̃w 1 + γ φ̃
Nsc krw −1 clw + Sh ρ̃h ṽs chw +
γ 1 + γη ∂ z̃ γ
à ! " à !à !8 #
1+γ ∂ 1+γ 1 + γ φ̃ ∂ p̃w l
Pe1 −Nsc krw c = 0 (10.26)
1−η ∂ x̃ γ 1 + γη ∂ x̃ w
(B.C.) at the seafloor is hydrostatic pressure, left and right side boundaries of the
2-D domain are assumed to be no-flow boundaries, whereas the bottom boundary
is maintained at hydrostatic pressure gradient. These initial and B.C.s are written
as:
ρw gD0 + ρw gz D0
I.C. : p̃w (z̃, x̃, 0) = = + z̃ (10.27)
ρw gLt Lt
203
D0
B.C.(1) : p̃w (0, x̃, t̃) = , (Hydrostatic pressure) (10.28)
Lt
∂ p̃w ∂ p̃w
B.C.(2) : (z̃, 0, t̃) = (z̃, Dx , t̃) = 0 , (No − flow) (10.29)
∂ x̃ ∂ x̃
∂ p̃w
B.C.(3) : (Dz , x̃, t̃) = 1 , (Hydrostatic gradient) (10.30)
∂ z̃
where D0 is seafloor depth, Dx is the width of the domain, and Dz is bottom of the
domain.
Sediment balance:
à !
∂ h i 1+γ ∂ h i
1 − φ̃ + Pe1 (1 − φ̃)ṽs = 0 (10.31)
∂ t̃ 1−η ∂ z̃
The initial velocity profile is obtained by assuming hydrostatic pressure, while the
1−η
I.C. : ṽs (z̃, x̃, 0) = , (for hydrostatic compaction) (10.32)
1 − φ̃
Organic balance:
à !
∂ h³ ´ i 1+γ ∂ h i ³ ´
1 − φ̃ α̃ + Pe1 (1 − φ̃)ṽs α̃ = −Da 1 − φ̃ α̃ (10.34)
∂ t̃ 1−η ∂ z̃
Initially, no organic carbon is present within the sediment, whereas the seafloor
204
Methane balance:
" # Ã !
∂ 1 + γ φ̃ ³ ´ 1+γ
Sw c̃lm + Sh c̃hm ρ̃h + Sg c̃gm ρ̃g + Pe1
∂ t̃ γ 1−η
"(Ã ! Ã !Ã !8 Ã !)
∂ 1 + γ φ̃ 1+γ 1 + γ φ̃ ∂ p̃w
Sw ṽs − Nsc krw −1 c̃lm +
∂ z̃ γ γ 1 + γη ∂ z̃
(Ã ! Ã !Ã !8 Ã !Ã !)
1 + γ φ̃ 1+γ 1 + γ φ̃ µw ∂ p̃g
Sg ṽs − Nsc krg − ρ̃g ρ̃g c̃gm +
γ γ 1 + γη µg ∂ z̃
à ! # à ! " à !à !8
1 + γ φ̃ 1+γ ∂ 1+γ 1 + γ φ̃ ∂ p̃w l
Sh c̃hm ρ̃h ṽs + Pe1 −Nsc krw c̃ −
γ 1−η ∂ x̃ γ 1 + γη ∂ x̃ m
à !à !8 à ! # " #
1+γ 1 + γ φ̃ µw ∂ p̃g g ∂ 1 + γ φ̃ ∂c̃lm
Nsc krg ρ̃g c̃m = Sw +
γ 1 + γη µg ∂ x̃ ∂ z̃ γ ∂ z̃
" #
∂ 1 + γ φ̃ ∂c̃lm MCH4
Sw + ρ̃s Da(1 − φ̃)α̃β (10.37)
∂ x̃ γ ∂ x̃ Morg
seafloor is set to zero, while methane concentration gradient is set to zero at the
The above equations (10.26, 10.31, 10.34 and 10.37) and boundary conditions
form a set of highly coupled, non-linear equations that are solved using a fully-
difficult to achieve using explicit schemes, especially during free gas migration and
non-hydrostatic compaction. The equations are expressed in residual form and the
in Appendix A5.
10.3 Results
The general dimensionless equations derived in the previous section are now
solved for different test cases. We start with simple, homogeneous, 2-D cases
and also simulate systems with combinations of fractures and heterogeneous per-
meability layers.
206
in free gas being advected with the sediment. However, free gas becomes mobile
if the critical gas saturation is exceeded. Mobile free gas can then rise buoyantly
and move as a separate phase. This will lead to the build-up of free gas beneath
the GHSZ, which may be sealed by the overlying low permeability hydrate layer.
other parameters remain same for both cases, with the following values: Pe1 =
geotherm of 0.04◦ C/m, seafloor depth of 2700 mbsl, and salinity equal to that of
standard seawater. Seawater density and viscosity are assumed to be 1030 kg/m3
and 1 cp, respectively, while gas density and viscosity are calculated at the local
sumed to be 0.02 MPa (Liu and Flemings, 2007), pore size distribution index (n) is
taken to be unity (Sun and Mohanty, 2006), and residual water saturation (Swr ) is
assumed to be 10% (Liu and Flemings, 2007). Residual gas saturation values in
the literature range from 4.5-17% (Schowalter, 1979) to more recent estimates of
2% (Moridis, 2003; Liu and Flemings, 2007) for gas hydrate systems. We choose
We now present simulation results for Nsc = 1000 and Nsc = 10. As shown
207
schematically in Chapter 9 (Figure 9.1), these two cases will explain how pore
connected free gas column. All simulation results are obtained for 2-D cases.
However, since these systems are homogeneous in the lateral direction, we explain
the results mostly using 1-D depth profiles. 2-D contour plots are shown for some
We first simulate the case Nsc = 1000, which indicates that sediment permeability
is large compared to the sedimentation rate. This will lead to overpressure due to
in this case will be caused by permeability reduction due to hydrate or free gas
build-up only. Figure 10.2 shows the pore pressure, gas pressure and lithostatic
Figure 10.3 shows the gas hydrate and free gas saturation at t̃ = 0.7. Compared
to gas saturation profiles shown previously, Figure 10.3 shows that gas in excess of
the critical saturation migrates upwards buoyantly to the base of the GHSZ. Thus,
maximum gas saturation occurs just below the GHSZ, while sediment sequences
far below the GHSZ contain free gas close to the residual saturation. Peak hydrate
saturation is about 25% at the base of the GHSZ, which reduces the absolute
sediment permeability, increases the gas capillary entry pressure and seals the
gas layer below. As more free gas accumulates beneath this hydrate layer, gas
208
0
Lithostatic
0.2 Gas pressure
Water pressure
0.4 Hydrostatic
0.6
Normalized depth
0.8
1.0
1.2
1.4
1.6
1.8
2.0
5 6 7 8 9 10
Normalized pressure
Figure 10.2: Lithostatic, gas pressure, water pressure, and hydrostatic pressure
profiles at time t̃ = 0.7 for Nsc = 1000. Gas pressure in excess of the water pressure
is the capillary pressure.
0
Gas hydrate
0.2 Free gas
0.4
0.6
Normalized depth
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0 0.1 0.2 0.3 0.4 0.5
Gas hydrate or free gas saturation
Figure 10.3: Gas hydrate and free gas saturation profiles at time t̃ = 0.7 for Nsc =
1000. Gas starts to migrate to the base of the GHSZ when it exceeds the critical
saturation.
209
pressure increases such that it exceeds the capillary entry pressure, causing free
gas to migrate into the GHSZ. This results in increased gas hydrate saturation at
Figure 10.4 and 10.5 show this transient behavior at time t̃ = 1. Water pressure
starts to deviate from the hydrostatic profile due to relatively large hydrate and free
gas saturations near the base of the GHSZ, which significantly reduce the relative
is seen just above the base of the GHSZ. This sharp increase is solely due to
gas migration from beneath the GHSZ and causes a longer connected free gas
column to build-up beneath the hydrate layer. At a later time, gas pressure exceeds
the new capillary entry pressure and again enters the GHSZ, causing an even
larger hydrate saturation at the base of the GHSZ. Thus, this process assumes
a cyclical nature, whereby gas keeps on accumulating beneath the hydrate layer,
which continually becomes more concentrated at the base of the GHSZ due to
periodical upward gas migration. However, as the gas column (and saturation)
builds up, gas pressure increases to a point where it exceeds the lithostatic stress.
At this point, the vertical gas effective stress (defined as the difference between
gas pressure and lithostatic stress) becomes zero, causing new or pre-existing
fractures to dilate, thereby allowing sudden free gas migration into the GHSZ or to
the ocean.
Following previous work (Flemings et al., 2003; Liu and Flemings, 2007), this
state at which gas pressure becomes equal to the overburden (lithostatic stress) is
termed ’critical-state’. This critical-state is shown in Figures 10.6, 10.7 and 10.8,
210
0
Lithostatic
0.2 Gas pressure
Water pressure
0.4 Hydrostatic
0.8
1.0
1.2
1.4
1.6
1.8
2.0
5 6 7 8 9 10
Normalized pressure
Figure 10.4: Lithostatic, gas pressure, water pressure, and hydrostatic pressure
profiles at time t̃ = 1 for Nsc = 1000.
0
Gas hydrate
0.2 Free gas
0.4
0.6
Normalized depth
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0 0.2 0.4 0.6 0.8 1
Gas hydrate or free gas saturation
Figure 10.5: Gas hydrate and free gas saturation profiles at time t̃ = 1 for Nsc =
1000.
211
which occurs at about t̃ ∼ 2.5. Figure 10.6 shows gas pressure just below the base
of the GHSZ become equal to the overburden, causing vertical gas effective stress
to be zero. Although we chose a relatively large value of Nsc for this simulation, the
water pressure profile reveals large overpressure beneath the GHSZ. This occurs
due to the low relative permeability of water caused by high hydrate and free gas
0
Lithostatic
0.2 Gas pressure
Water pressure
0.4 Hydrostatic
0.6
Normalized depth
0.8
1.0
1.2
1.4
1.6
1.8
2.0
5 6 7 8 9 10
Normalized pressure
Figure 10.6: Lithostatic, gas pressure, water pressure, and hydrostatic pressure
profiles for Nsc = 1000 when gas pressure becomes equal to the overburden at the
critical-state (t̃ ∼ 2.5).
Hydrate and free gas saturation profiles at this critical-state are shown in Figure
10.7. Compared to previous transient results, this plot shows a much higher hy-
drate saturation at the base of the GHSZ, which is required to trap a much longer
connected free gas column beneath the hydrate layer. The non-dimensional thick-
212
0
Gas hydrate
0.2 Free gas
0.4
0.6
Normalized depth
0.8
1.0
1.2
1.4
1.6
1.8
2
0 0.2 0.4 0.6 0.8 1
Gas hydrate or free gas saturation
Figure 10.7: Gas hydrate and free gas saturation profiles at the critical-state for
Nsc = 1000. A deep connected free gas column forms beneath the GHSZ and is
trapped by the high saturation hydrate layer.
0.6
Normalized depth
0.5 0.5
0.4
1
0.3
0.2
1.5
0.1
0 0.5 1 1.5 2
Normalized lateral distance
Free gas saturation
0.7
Normalized depth
0.6
0.5
0.5
0.4
1
0.3
0.2
1.5
0.1
0
0 0.5 1 1.5 2
Normalized lateral distance
Figure 10.8: 2-D contour plot of hydrate and free gas saturations at the critical-state
for Nsc = 1000.
213
ness of the free gas column from Figure 10.7 is about 0.4, which multiplied by the
depth to the base of the GHSZ (∼450 mbsf for Blake Ridge type seafloor condi-
tions) yields a gas column about 180 meters thick. Finally, Figure 10.8 shows the
2-D contour plot of hydrate and free gas saturation at this critical-state.
Case 2: Nsc = 10
Lowering Nsc from 1000 to 10 causes sediment permeability to decrease with re-
spect to the sedimentation rate. This causes overpressure to develop in the system
pressure becomes elevated even without any contribution from permeability reduc-
tion due to hydrate or free gas. Intuitively, this suggests that such systems would
allow a relatively shorter gas column to develop before gas pressure exceeds the
profiles when gas pressure becomes equal to the overburden for Nsc = 10 (critical-
state), at time t̃ ∼ 1.5. Compared to Figure 10.6 (Nsc = 1000), pore pressure
sequence. Because of this already elevated background water pressure, gas pres-
sure does not has to increase much to exceed the overburden. Consequently, only
a relatively thin free gas column develops before gas effective stress goes to zero,
Figure 10.10 shows the gas hydrate and free gas saturation profiles at this
critical-state for Nsc = 10. Transient evolution of hydrate and free gas follows a pat-
214
0
Lithostatic
0.2 Gas pressure
Water pressure
0.4 Hydrostatic
0.6
Normalized depth
0.8
1.0
1.2
1.4
1.6
1.8
2.0
5 6 7 8 9 10
Normalized pressure
Figure 10.9: Lithostatic, gas pressure, water pressure, and hydrostatic pressure
profiles for Nsc = 10 when gas pressure becomes equal to the overburden at the
critical-state (t̃ ∼ 1.5).
tern similar to the previous case, whereby gas migrates buoyantly to the base of the
GHSZ, causing a local spike in hydrate saturation at the base, which seals further
gas migration. Since the gas saturation profile shows a relatively thin connected
column, the peak hydrate saturation needed at the base to seal this column is also
relatively low (∼60%), compared to the 80% peak saturation in Figure 10.7. The
dimensionless thickness of the free gas column in this case is about 0.1, which
thickness of about 45 m.
Thus, the two cases shown above demonstrate that sediment permeability with
0
Gas hydrate
0.2 Free gas
0.4
0.6
Normalized depth
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0 0.2 0.4 0.6 0.8 1
Gas hydrate or free gas saturation
Figure 10.10: Gas hydrate and free gas saturation profiles at the critical-state for
Nsc = 10 (t̃ ∼ 1.5). Compared to the thick gas column in Figure 10.7, only a
relatively thin free gas column forms beneath the GHSZ in this case before gas
pressure becomes equal to the overburden.
216
in determining the thickness of the connected free gas column beneath the GHSZ.
Therefore, gas hydrate sites characterized by higher values of Nsc could have better
Fracture networks can dominate fluid flow in gas hydrate systems, serving as high
permeability conduits and causing localized high concentration of hydrate and free
gas around these networks (Weinberger and Brown, 2006). Such networks might
where gas hydrate is distributed very heterogeneously across varying length scales
(Tréhu et al., 2004; Weinberger and Brown, 2006). Although several past studies
the current numerical models only simulate simple 1-D homogeneous sediment
sequences. In comparison, our 2-D model offers the capability to assign different
rock properties to different lithologies (such as sand, clay, fractures, etc.), thereby
tures to move down with the sediment. We start with the case where a single
vertical fracture is introduced in the system and follow transient gas hydrate and
free gas accumulation as this fracture moves down with the sediment. Due to this
between layers of different permeabilities are tracked in time. By knowing the inter-
face position within any given gridblock, the horizontal and vertical permeabilities
simulated.
Figure 10.11. The vertical fracture has permeability 100 times higher than the
surrounding sediment, i.e. Nsc = 20 for the sediment matrix and Nsc = 2000 for
the fracture. The primary transport parameters are assigned the following values:
properties of water, hydrate and gas remain the same as stated in section 10.3.1.
BHSZ BHSZ
Figure 10.11: Permeability map showing initial location of a single, high perme-
ability vertical fracture with permeability 100 times higher than the surrounding
sediment.
218
Simulation results are now presented at different transient states. We first run
t̃ = 0.5. At this time (t̃ = 0.5) the fracture shown in Figure 10.11 is introduced.
The system response to this fracture is now studied at later time values. Figure
10.12 shows the gas hydrate saturation contours at time t̃ = 0.6 for this single
vertical fracture system. The position of the fracture in this contour plot, and in
subsequent plots, is shown through the set of dashed lines. The effect of the frac-
ture in focusing flow along this high permeability conduit is clearly seen through
the elevated hydrate saturations along and around the fracture. Peak gas hydrate
saturation within the fracture is about 15%, while peak hydrate saturation in the
surrounding sediments is about 7%. Free gas saturation contours at time t̃ = 0.6,
shown in 10.13, also depict maximum gas saturation along the fracture. However,
free gas saturation at this time is too low to cause any lateral migration.
Figure 10.14 shows gas hydrate saturation contours at a later time (t̃ = 1.0).
Compared to Figure 10.12, peak hydrate saturation increases to ∼20% and occurs
within the fracture just above the base of the GHSZ. This peak value is about twice
the peak saturation in the surrounding sediments at the base of the GHSZ (∼10%).
As the fracture moves away from the seafloor, the shallower sediments are restored
to the original lower permeability, causing them to have hydrate saturation equal to
the surrounding sediments at similar depth. Figure 10.15 shows free gas saturation
contours at t̃ = 1.0. Buoyant free gas migrates upwards and gets sealed by the low
permeability hydrate layer at the base of the GHSZ. Maximum free gas saturation
occurs just below the GHSZ along the high permeability fracture (∼50%). At this
219
0.2 0.14
0.4
0.12
0.6
Normalized depth 0.1
0.8
0.08
1
1.2 0.06
1.4
0.04
1.6
0.02
1.8
0.5 1 1.5
Normalized lateral distance
Figure 10.12: Gas hydrate saturation contours at time t̃ = 0.6 for the vertical frac-
ture system shown schematically in Figure 10.11. Dashed lines denote the position
of the fracture.
0.2 0.04
0.4 0.035
0.6
Normalized depth
0.03
0.8
0.025
1
0.02
1.2
0.015
1.4
0.01
1.6
0.005
1.8
0.5 1 1.5
Normalized lateral distance
Figure 10.13: Free gas saturation contours at time t̃ = 0.6 for the vertical fracture
system shown schematically in Figure 10.11. Dashed lines denote the position of
the fracture.
220
0.2
0.2 0.18
0.4 0.16
0.6 0.14
Normalized depth
0.8 0.12
1 0.1
1.2 0.08
0.06
1.4
0.04
1.6
0.02
1.8
0
0.5 1 1.5
Normalized lateral distance
Figure 10.14: Gas hydrate saturation contours at time t̃ = 1.0 for the vertical frac-
ture system shown schematically in Figure 10.11. Peak gas hydrate saturation
occurs at the base of the GHSZ within the fracture.
0.2 0.45
0.4
0.4
0.35
0.6
Normalized depth
0.3
0.8
0.25
1
0.2
1.2
0.15
1.4
0.1
1.6
0.05
1.8
0
0.5 1 1.5
Normalized lateral distance
Figure 10.15: Free gas saturation contours at time t̃ = 1.0 for the vertical frac-
ture system shown schematically in Figure 10.11. Peak gas saturation is about
50% within the fracture, while lateral gas migration from the fracture causes gas
saturation in neighboring gridblocks to increase to about 30%.
221
high saturation, free gas also migrates laterally, causing gridblocks just below the
GHSZ and neighboring the fracture to also have relatively higher saturations.
Finally, we show gas hydrate saturation contours at time t̃ = 1.5 in Figure 10.16,
when the fracture has almost passed the GHSZ. The top of the fracture is very
close to the base of the GHSZ, causing a relatively small increase in gas hydrate
saturation within the fracture column. Peak hydrate saturation of about 12% within
the fracture is marginally greater than the 10% saturation in sediments close to
the base of the GHSZ but far away from the fracture. Overall, as the fracture
moves out of the GHSZ, gas hydrate distribution becomes much more uniform
along the lateral direction. Free gas saturation below the GHSZ at t̃ = 1.5 (Figure
10.17) increases to a peak value of about 60% within the fracture and spreads
out laterally away from the fracture. Similar to the region above the base of the
GHSZ, free gas saturation also becomes more laterally uniform as the fracture is
Thus, presence of fractures can significantly affect gas hydrate and free gas
distribution by focusing fluid flow along these conduits. This effect, however, is dy-
namic in nature. As these fractures move out of the GHSZ into deeper sediments,
Apart from vertical fractures, our model can also incorporate stratigraphy of vary-
ing permeabilities to simulate horizontal or dipping sand layers between low per-
meability clay layers. We first show pore pressure evolution for simple horizontal
222
0.11
0.2
0.1
0.4
0.09
0.6
Normalized depth
0.08
0.8 0.07
0.06
1
0.05
1.2
0.04
1.4
0.03
1.6 0.02
1.8 0.01
0.5 1 1.5
Normalized lateral distance
Figure 10.16: Gas hydrate saturation contours at time t̃ = 1.5 for the vertical frac-
ture system (Figure 10.11). Peak gas hydrate saturation within the fracture is al-
most equal to the value in surrounding sediment as the fracture moves out of the
GHSZ.
beds and then move to more realistic cases of dipping sand layers.
We start with the simplest case of a horizontal high permeability sand layer de-
posited in low permeability clay sediments. This example is used to illustrate the
downward movement of the high permeability sand layer and the corresponding
pore pressure evolution. The parameter Nsc is 10 for the low permeability sedi-
ments and 1000 for the sand layer, i.e., the sand layer is 100 times more perme-
able than the surrounding sediments. Following values are assigned to the other
0.2
0.4 0.5
0.6
Normalized depth
0.4
0.8
1 0.3
1.2
0.2
1.4
1.6 0.1
1.8
0
0.5 1 1.5
Normalized lateral distance
Figure 10.17: Free gas saturation contours at time t̃ = 1.5 for the vertical fracture
system shown schematically in Figure 10.11. Peak gas saturation is 67% within
the fracture, while lateral gas migration from the fracture causes gas saturation in
neighboring gridblocks to increase to about 40%.
critical gas saturation of 5%. Seafloor conditions and other parameters pertain-
We start the simulation at time t̃ = 0 with low permeability sediments and de-
posit the sand layer uniformly across the lateral section from time t̃ = 1.0 to 1.5.
Figure 10.18 shows various pressure profiles along the normalized depth at di-
mensionless times t̃ = 1.75 and 2.25. The sand interval is marked by dotted lines
in both plots. At time t̃ = 1.75, the sand layer starts extends from z̃ = 0.25 to z̃ = 0.6
and the pore pressure gradient within this interval is almost hydrostatic due to its
higher absolute permeability. Conversely, pore pressure gradient above and below
224
0.6 0.6
Normalized depth
Normalized depth
0.8 0.8 Sand layer
1 1
1.2 1.2
1.4 1.4
1.6 1.6
1.8 1.8
2 2
6 7 8 9 6 7 8 9
Normalized pressure Normalized pressure
Figure 10.18: Pressure evolution for the case of a sand layer (high permeability)
deposited in low permeability sediments uniformly across the lateral section. Sand
layer is deposited from time t̃ = 1.0 to 1.5. (a) Pore pressure, gas pressure and
lithostatic stress profiles along depth at time t̃ = 1.75. The dotted interval repre-
sents the sand layer and is characterized by pore pressure gradient that is close to
the hydrostatic gradient. (b) At at later time, t̃ = 2.25, the sand layer moves deeper
and shrinks in thickness due to compaction.
225
the sand layer is greater than the hydrostatic gradient due to its lower absolute
is more clearly seen at time t̃ = 2.25, when the sand layer moves further down into
the sediment column and also shrinks in thickness due to compaction and porosity
loss. This feature characterizes all simulations shown later, where the thickness
0.15
Normalized depth
0.5
0.1
1
0.05
1.5
0
1 2 3 4 5 6 7 8 9
Normalized lateral distance
Free gas saturation
0.25
Normalized depth
0.5 0.2
0.15
1
0.1
1.5 0.05
0
1 2 3 4 5 6 7 8 9
Normalized lateral distance
Figure 10.19: Gas hydrate and free gas saturation at time t̃ = 2.25. Since the high
permeability sand layer extends uniformly across the lateral section, hydrate and
free gas saturation are also laterally homogeneous.
of the sand layer continuously decreases as it loses porosity and is buried deeper
Gas hydrate and free gas saturation contours at t̃ = 2.25 are shown in Figure
10.19. Due to the laterally uniform deposition of the sand layer, hydrate and free
gas saturation also remain constant across the lateral section. Thus, this system
226
is essentially 1-D in nature and since the lateral ends of the domain are no-flow
no fluid focusing or enhanced concentration of hydrate or free gas within the high
model dipping sand layers as well as combinations of fractures and sand layers.
To model a dipping sand layer, we start with a pre-existing sand layer at a given dip
angle within the sediment and deposit low permeability clay on it. The downward
movement of this sand layer and the corresponding transient hydrate/free gas evo-
lution is then tracked through time. The permeability map, illustrated schematically
in Figure 10.20, shows the initial position of of this sand layer. Similar to the case of
the horizontal layer discussed above, the sand layer is assigned an absolute per-
meability 100 times greater than the surrounding sediments. The physical domain
for all simulations in this section is z̃ ∈ [0, 2] and x̃ ∈ [0, 10]. The parameter Nsc is
10 for the low permeability sediments and 1000 for the sand layer. The following
values are assigned to the other transport parameters: Pe1 = 0.1, Da = 10, β = 6,
physical properties of water, hydrate and gas remain the same as stated in section
10.3.1. We now show results for two cases with different dip angles.
Case 1 - Low dip angle: In the first case, we start with a low dip angle of about
BHSZ BHSZ
Figure 10.20: Schematic representation of the high permeability sand layer sur-
rounded by low permeability clay sediments.
sediment matrix and start depositing low permeability clay at the seafloor at time
t̃ = 0. Gas hydrate saturation contours at time t̃ = 0.25 are shown in Figure 10.21.
Compared to the case of a horizontal layer, this plot shows significant concentration
of hydrate within the sand layer. Peak gas hydrate saturation within the sand layer
increases to about 30% but the color axis is scaled to a maximum of 15% to show
the contours in the rest of the domain more clearly. However, the contour plot is not
very convenient to look at because of the x-axis range being five time the z-axis
range in Figure 10.21. Thus, gas hydrate and free gas saturations at time t̃ = 0.25
are replotted in Figure 10.22 with the y-axis now rescaled to represent a vertical
The effect of fluid focusing through the high permeability sand layer is evident
in the contour plots in Figure 10.22. As with Figure 10.21, the color axis for the
228
0.1
Normalized depth
1.0
2.0
0 Normalized lateral distance 10.0 0.05
Figure 10.21: Gas hydrate saturation contours at t̃ = 0.25 for the low dip angle with
”true” axis scale, i.e., no vertical exaggeration.
229
to show the distribution in the rest of the GHSZ more clear, which otherwise gets
dominated by the high saturation gridblocks only. The dashed lines in both contour
plots denote the position of the sand layer that has moved away from the seafloor
due to sedimentation. Peak gas hydrate saturation is about 30% within the sand
layer and about 8% in the region close to the base of the GHSZ outside the sand
layer. Free gas is also focused within the sand layer with saturations as high as
50%. Such high gas saturations cause it to migrate laterally out of the sand layer
and into the low permeability sediments just below the GHSZ.
Figure 10.22: Gas hydrate and free gas saturation contours at t̃ = 0.25 with a
vertical exaggeration (VE) of about 2:1. Dashed lines denote the position of the
sand layer within the low permeability sediment matrix.
230
At a later time, t̃ = 0.75, the sand layer is buried even deeper into the sediments
(Figure 10.23). Similar to the case of the fracture moving down (section 10.3.2), as
the sand layer moves out of the GHSZ, its effect on gas hydrate saturation becomes
diminished. Figure 10.23 shows the saturation contours for this case, where only
a small section of the sand layer is present within the GHSZ. This small region
within the sand layer is close to the base of the GHSZ and has hydrate saturation
of about 12%, compared to about 8% saturation near the base on either side of the
sand layer. Free gas saturation increases to about 67% within the sand layer just
below the GHSZ with neighboring sediments outside the sand layer at about 30%.
At an even later time, the sand layer moves completely out of the system causing
We next simulate a case with a dip of about 4 degrees to study the effect of
the dip angle on hydrate saturation. All other parameters remain the same as the
previous case. Compared to the previous case, which had a dip of about 2 degrees,
Figure 10.24 shows much higher saturations within the sand layer at time t̃ = 0.3.
Peak hydrate saturation within the sand layer is about 45%, significantly higher than
the 8% peak hydrate saturation in the low permeability sediments within the GHSZ.
Free gas saturation is about 66% within the sand layer just below the GHSZ and,
similar to previous figures, spreads laterally into the low permeability sediments
outside the sand layer. Thus, keeping all other parameters same, higher dip angle
leads to relatively higher fluid focusing and hydrate saturation within the GHSZ.
231
Figure 10.23: Gas hydrate and free gas saturation contours at t̃ = 0.75 with a
vertical exaggeration (VE) of about 2:1. Dashed lines denote the position of the
sand layer within the low permeability sediment matrix.
232
Figure 10.24: Gas hydrate and free gas saturation contours at t̃ = 0.3 for the case
of higher dip angle and with a vertical exaggeration (VE) of about 2:1. Color axis
for the hydrate saturation contour plot is scaled to a maximum of 15%. Dashed
lines denote the position of the sand layer within the low permeability sediment
matrix.
233
After studying fractures and dipping sand layers in the previous two sections,
we now combine these two features to simulate a more general setting where a
fracture intersects a dipping sand layer. Following parameter values are used:
ities, capillary pressure, and physical properties of water, hydrate and gas remain
the same as stated in section 10.3.1. The initial system consists of a fracture
located around the center of the grid and a dipping sand layer with a dip angle of
about 2 degrees. Both the fracture and the sand layer have absolute permeabilities
Figure 10.25 shows gas hydrate and free gas saturation contours for this sys-
tem after dimensionless time t̃ = 0.25. This figure also has a VE of about 2:1, with
the dashed lines indicating the location of the fracture and the dipping sand layer.
It can be seen that gas hydrate is most concentrated near the intersection of the
sand layer with the fracture, which has a peak saturation of about 48%. However,
as before, the color axis for the hydrate saturation plot is scaled to a maximum of
about 15% to show hydrate distribution in the rest of the GHSZ more clearly. Free
gas saturation is maximum within the fracture, with peak saturation of 67% just
below the GHSZ and about 40% within the sand layer around the region x̃ = 0.
At a later time, t̃ = 0.75, a significant portion of the sand layer moves out of the
GHSZ, causing hydrate distribution to be more uniform laterally within the GHSZ
234
Figure 10.25: Gas hydrate and free gas saturation contours at t̃ = 0.25 for the case
of combined fracture and lower dip angle sand layer, with a vertical exaggeration
(VE) of about 2:1. Color axis for the hydrate saturation contour plot is scaled to a
maximum of 15%. Dashed lines denote the position of the fracture and the sand
layer within the low permeability sediment matrix.
235
Figure 10.26: Gas hydrate and free gas saturation contours at t̃ = 0.75 for the case
of combined fracture and lower dip angle sand layer, with a vertical exaggeration
(VE) of about 2:1. Dashed lines denote the position of the fracture and the sand
layer within the low permeability sediment matrix.
236
(Figure 10.26). A small portion of the fracture still remains within the GHSZ, which
causes a small increase in hydrate saturation (∼12%) around that region. Free gas
saturation is now about greater than 65% in a wide region within the sand layer and
beneath the GHSZ. Both the fracture and the sand layer act as conduits to focus
free gas within these more permeable sediments, which then rises buoyantly and
The simulation results presented above show that lithology plays an important
lated relatively simple systems with single fractures and/or dipping sand layers,
terms of fracture networks, multiple sand layers or isolated sand lenses. These
the numerical model can be used in the future to simulate systems with consider-
able heterogeneity that mimic natural gas hydrate systems more accurately.
10.4 Conclusions
chapter to simulate gas hydrate and free gas accumulation in marine sediments
sediment heterogeneity and lateral fluid flow in the system. Following conclusions
• Free gas migrates buoyantly upwards into the GHSZ when its critical gas sat-
237
uration is exceeded. As free gas enters the GHSZ, gas hydrate precipitates
at the base, causing an increase in the capillary entry pressure for the free
gas. This restricts further gas entry into the GHSZ. Thus, free gas starts to
accumulate beneath the GHSZ into a connected gas column. The thickness
group Nsc . Relatively higher values of Nsc lead to lower overpressure and
development of a long gas column before gas pressure at the base of the
GHSZ exceeds the lithostatic stress. At this point, fractures can dilate and
vent the free gas below the hydrate, but we do not model this process in our
simulations, i.e., our simulations are stopped whenever vertical gas effective
stress becomes zero. Conversely, relatively lower values of Nsc leads to con-
gas column, i.e., fracturing can occur for a short gas column.
• Vertical fractures in the sediments focus fluid flow along this high permeability
conduit causing relatively higher gas hydrate and free gas saturations within
deeper into the sediments, hydrate saturation becomes more uniform later-
ally.
sediments are also simulated using our numerical model. Hydrate and free
gas are concentrated within this sand layer due to fluid focusing. Higher dip
angle leads to greater hydrate and free gas saturations within these sand
238
layers. However, similar to the case of fractures, the effect of these sand
Chapter 11
11.1 Conclusions
General numerical models for simulating gas hydrate accumulation in marine sedi-
ments were developed in this thesis. In this chapter, we first summarize the conclu-
sions from these models/theory and then present some future research directions.
The overall conclusions are divided into the following three categories:
sionless groups controlling gas hydrate saturation at steady state are the
two Peclet numbers (Pe1 and Pe2 ), the Damkohler number (Da), and the nor-
malized organic carbon input at the seafloor (β). The type of gas hydrate
accumulation, i.e., no hydrate and hydrate with or without free gas below,
average gas hydrate saturation is a function of Pe1 and the organic carbon
converted within the GHSZ. For deeper methane sources, average hydrate
saturation depends on the net fluid flux (Pe1 + Pe2 ) and Pe1 . Further, settings
dominated by in situ methane can be divided into diffusive and advective sys-
tems, with increase in sedimentation rates (Pe1 ) causing higher hydrate sat-
240
urations in the former regime and lower hydrate saturation in the latter. For
high fluid flux (Pe1 + Pe2 ) and low sedimentation rates (Pe1 ). Studying gas hy-
sulfate profile and depth of the SMT below the seafloor are shown to be func-
tions of the net methane flux from depth for gas hydrate systems dominated
by deeper sources. Since gas hydrate saturation and thickness of the gas
hydrate layer within the GHSZ are also functions of net methane flux for this
type of methane source, the scaled depth of the SMT becomes a fast and in-
expensive proxy for quantifying hydrate saturation. Shallow SMT depths are
shown through modeling and analytical theory to be indicative of high gas hy-
drate saturation. The depth to the first occurrence of gas hydrate below the
seafloor is shown to be about 10-12 times the depth of the SMT for most gas
hydrate settings. Finally, application of this new method to several sites along
Cascadia Margin reveals a good match with saturations estimated from re-
faster sediment velocities through the GHSZ, causing a net decrease in the
for our set of boundary conditions. The depth of the GHSZ below the seafloor
small even for large overpressures. Overpressure also limits the thickness of
the connected gas column beneath the GHSZ. This scenario is demonstrated
by allowing free gas to migrate upwards buoyantly as soon as its critical gas
tion at the base of the GHSZ, which restricts further gas migration into the
GHSZ. The development of this free gas column is tracked through transient
simulations, which show that high values of Nsc lead to deep connected gas
columns, while low Nsc values yield shorter gas columns before fracturing
occurs. Finally, several different cases are simulated that show the effect of
fluid focusing within high permeability sediments. These include gas hydrate
systems with vertical fractures, dipping sand layers and combinations of both.
In general, these simulations show that hydrate as well as free gas become
This thesis incorporates most of the primary factors believed to control gas hy-
drate and free gas formation and distribution in marine sediments. However, some
drate formation and dissociation affects the pore water salinity, which, in turn,
changes the phase equilibrium of the system (Zatsepina and Buffett, 1998;
Milkov et al., 2004; Liu and Flemings, 2006). A simple chloride balance was
on pore water salinity and its role as a gas hydrate proxy. However, the salin-
ity change was not coupled to the thermodynamic and transport equations,
so that changing salinity did not affect hydrate phase equilibrium. Although,
for general cases, the change in salinity and its effect on phase equilibrium
appears to be small from these preliminary simulations, such effects may be-
come dominant in special sites characterized by high gas flux that can cause
massive hydrate formation and large salinity anomalies (Milkov et al., 2004;
• Compositional effect on gas hydrate and free gas distribution: All sim-
ulations and theory presented in this thesis assume methane to be the only
hydrate former. However, at several gas hydrate settings (e.g., Gulf of Mex-
ico) higher alkanes like ethane and propane can be present in amounts large
(Milkov and Sassen, 2000). In such cases, gas hydrate may form sI or sII
structures that can coexist with free gas over finite depths instead of a sharp
boundary between hydrate and gas for methane hydrates. Thus, BSRs might
not be easily seen on seismic profiles for such settings. Hence, the mod-
• Blanking effect and chaotic zones: Amplitude blanking refers to the at-
Dillon, 2001). In other words, gas hydrate preferentially forms in high poros-
contrast between different lithologies can be tested with our 2-D model in
blanking, chaotic zones may develop due to migration of free gas into the
form as a result of high permeability conduits and focused fluid flow. Some
of the simple heterogeneous features are studied in this thesis, e.g., faults
Apart from the stratigraphic features, the potential of structural traps in focus-
Bibliography
Adisasmito, S., Frank, R. J., and Sloan, E. D., J. (1991). Hydrates of carbon dioxide
and methane mixtures. J. Chem. Eng. Data, 36:68–71.
Bear, J. (1988). Dynamics of Fluids in Porous Media. New York : Dover Publica-
tions Inc.
Bhatnagar, G., Chapman, W. G., Dickens, G. R., Dugan, B., and Hirasaki, G. J.
(2007). Generalization of gas hydrate distribution and saturation in marine sed-
iments by scaling of thermodynamic and transport processes. Am. J. Sci.,
307:861–900.
Bhatnagar, G., Chapman, W. G., Dickens, G. R., Dugan, B., and Hirasaki,
G. J. (2008). Sulfate-methane transition as a proxy for average methane
hydrate saturation in marine sediments. Geophys. Res. Lett., 35, L03611,
doi:10.1029/2007GL032500.
Borowski, W. S., Paull, C. K., and Ussler III, W. (1996). Marine pore-water sul-
fate profiles indicate in situ methane flux from underlying gas hydrate. Geology,
24(7):655–658.
Borowski, W. S., Paull, C. K., and Ussler III, W. (1999). Global and local varia-
tions of interstitial sulfate gradients in deep-water, continental margin sediments:
Sensitivity to underlying methane and gas hydrates. Mar. Geol., 159:131–154.
Boudreau, B. P. and Westrich, J. T. (1984). The dependence of bacterial sulfate
reduction on sulfate concentration in marine sediments. Geochim. Cosmochim.
Acta, 48:2503–2516.
Buffett, B. A. and Archer, D. (2004). Global inventory of methane clathrate: Sensi-
tivity to changes in the deep ocean. Earth Planet. Sci. Lett., 227:185–199.
Cao, Z. T., Tester, J. W., and Trout, B. L. (2002). Sensitivity analysis of hydrate ther-
modynamic reference properties using experimental data and ab initio methods.
J. Phys. Chem. B, 106(31):7681–7687.
Castellini, D. G., Dickens, G. R., Snyder, G. T., and Ruppel, C. D. (2006). Barium
recycling in shallow sediment above active mud volcanoes in the Gulf of Mexico.
Chem. Geol., 226:1–30.
Claypool, G. E. and Kaplan, I. R. (1974). Natural Gases in Marine Sediments,
chapter Methane in marine sediments, pages 99–139. Plenum Press. New York.
Claypool, G. E. and Kvenvolden, K. A. (1983). Methane and other hydrocarbon
gases in marine sediment. Annu. Rev. Earth Planet. Sci., 11:299–327.
Clennell, M. B., Hovland, M., Booth, J. S., Henry, P., and Winters, W. J. (1999).
Formation of natural gas hydrates in marine sediments 1. Conceptual model of
gas hydrate growth conditioned by host sediment properties. J. Geophys. Res.,
104:22985–23003.
Coffin, R., Pohlman, J., Gardner, J., Downer, R., Wood, W., Hamdan, L., Walker,
S., Plummer, R., Gettrust, J., and Diaz, J. (2007). Methane hydrate exploration
on the mid Chilean coast: A geochemical and geophysical survey. J. Petrol. Sci.
Eng., 56(1):32–41.
Collett, T. S. (2000). Quantitative well-log analysis of in-situ natural gas hydrates.
PhD thesis, Colorado School of Mines, Golden, CO.
Collett, T. S. (2002). Energy resource potential of natural gas hydrates. AAPG
Bull., 86:1971–1992.
Collett, T. S. and Ladd, J. (2000). Detection of gas hydrate with downhole logs and
assessment of gas hydrate concentrations (saturations) and gas volumes on the
Blake Ridge with electrical resistivity log data. In Paull, C. K., Matsumoto, R.,
Wallace, P. J., and Dillon, W. P., editors, Proc. ODP, Sci. Results, volume 164,
pages 179–191, College Station, TX. Ocean Drilling Program.
247
Davie, M. K. and Buffett, B. A. (2001). A numerical model for the formation of gas
hydrate below the seafloor. J. Geophys. Res., 106:497–514.
Davie, M. K. and Buffett, B. A. (2003a). Sources of methane for marine gas hydrate:
inferences from a comparison of observations and numerical models. Earth
Planet. Sci. Lett., 206:51–63.
Davie, M. K. and Buffett, B. A. (2003b). A steady state model for marine hydrate
formation: Constraints on methane supplpy from pore water sulfate profiles. J.
Geophys. Res., 108, 2495, doi:10.1029/2002JB002300.
Davie, M. K., Zatsepina, O. Y., and Buffett, B. A. (2004). Methane solubility in
marine hydrate environments. Mar. Geol., 203:177–184.
de Caritat, P. P. (1989). Note on the maximum upward migration of pore water in
response to sediment compaction. Sediment. Geol., 65:371–377.
de Roo, J. L., Peters, C. J., Lichtenthaler, R. N., and Diepen, G. A. M. (1983). Oc-
currence of methane hydrate in saturated and unsaturated solutions of sodium
chloride and water in dependence of temperature and pressure. AICHE J.,
29:651–657.
D’Hondt, S. L., Jørgensen, B. B., and Miller, et al., D. J., editors (2003). Proc. ODP,
Init. Repts., volume 201, College Station, TX. Ocean Drilling Program.
Dickens, G. R. (2001). The potential volume of oceanic methane hydrates with
variable external conditions. Org. Geochem., 32:11791193.
Dickens, G. R. (2003). Rethinking the global carbon cycle with a large, dy-
namic and microbially mediated gas hydrate capacitor. Earth Planet. Sci. Lett.,
213:169–183.
Dickens, G. R., Castillo, M. M., and Walker, J. C. G. (1997). A blast of gas in the
latest Paleocene: Simulating first order effects of massive dissociation of oceanic
methane hydrate. Geology, 25:259–262.
Dickens, G. R., O’Neil, J. R., Rea, D. K., and Owen, R. M. (1995). Dissociation of
oceanic methane hydrate as a cause of the carbon isotope excursion at the end
of the Paleocene. Paleoceanography, 10:965–971.
Dickens, G. R. and Quinby-Hunt, M. S. (1994). Methane hydrate stability in sea-
water. Geophys. Res. Lett., 21:2115–2118.
Duan, Z., Moller, N., Greenberg, J., and Weare, J. H. (1992). The prediction of
methane solubility in natural waters to high ionic strength from 0 to 250 c and
from 0 to 1600 bar. Geochim. Cosmochim. Acta, 56:1451–1460.
Dugan, B. and Flemings, P. B. (2000). Overpressure and fluid flow in the New
Jersey continental slope: implications for slope failure and cold seeps. Science,
289:288–291.
248
Falta, R. W., Pruess, K., Javandel, I., and Witherspoon, P. A. (1992). Numerical
modeling of steam injection for the removal of nonaqueous phase liquids from
the subsurface 1. Numerical formulation. Water Resour. Res., 28(2):433–449.
Finkbeiner, T., Zoback, M., Flemings, P., and Stump, B. (2001). Stress, pore pres-
sure, and dynamically constrained hydrocarbon columns in the South Eugene
Island 330 field, northern Gulf of Mexico. AAPG Bull., 85(6):1007–1031.
Flemings, P. B., Liu, X., and Winters, W. (2003). Critical pressure and multiphase
flow in Blake Ridge hydrates. Geology, 31:1057–1060.
Floodgate, G. D. and Judd, A. G. (1992). The origins of shallow gas. Cont. Shelf
Res., 12:1145–1156.
Haeckel, M., Suess, E., Wallman, K., and Rickert, D. (2004). Rising methane gas
bubbles form massive hydrate layers at the seafloor. Geochim. Cosmochim.
Acta, 68:4335–4345.
Henry, P., Thomas, M., and Clennell, M. B. (1999). Formation of natural gas hy-
drates in marine sediments, 2, Thermodynamic calculations of stability condi-
tions in porous sediments. J. Geophys. Res., 104:2300523022.
Hesse, R., Frape, S. K., Egeberg, P. K., and Matsumoto, R. (2000). Stable isotope
studies (Cl, O, and H) of interstitial waters from Site 997, Blake Ridge gas hydrate
field, West Atlantic. In Paull, C. K., Matsumoto, R., Wallace, P. J., and Dillon,
W. P., editors, Proc. ODP, Sci. Results, volume 164, pages 129–137, College
Station, TX. Ocean Drilling Program.
Holbrook, W. S., Hoskins, H., Wood, W. T., Stephen, R. A., Lizzaralde, D., and Leg
164 Science Party (1996). Methane hydrate and free gas on the Blake Ridge
from vertical seismic profiling. Science, 273:1840–1843.
Holder, G. D., G., C., and Papadopoulos, K. D. (1980). Thermodynamic and molec-
ular properties of gas hydrates from mixtures containing methane, argon and
krypton. Ind. Eng. Chem. Fundam., 19:282–286.
Hornbach, M. J., Holbrook, W. S., Gorman, A. R., Hackwith, K. L., Lizzaralde, D.,
and Pecher, I. (2003). Direct seismic detection of methane hydrate on the Blake
Ridge. Geophysics, 68:92–100.
Hyndman, R. D., Yuan, T., and Moran, K. (1999). The concentration of deep sea
gas hydrates from downhole electrical resistivity logs and laboratory data. Earth
Planet. Sci. Lett., 172:167–177.
250
Joye, S. B., Boetius, A., Orcutt, B. N., Montoya, J. P., Schulz, H. N., Erickson,
M. J., and Lugo, S. K. (2004). The anaerobic oxidation of methane and sulfate
reduction in sediments from Gulf of Mexico cold seeps. Chem. Geol., 205:219–
238.
Kastner, M., Kvenvolden, K. A., Whiticar, M. J., Camerlenghi, A., and Lorenson,
T. D. (1995). Relation between pore fluid chemistry and gas hydrates associated
with bottom-simulating reflectors at the Cascadia Margin, Sites 889 and 892. In
Carson, B., Westbrook, G. K., Musgrave, R. J., and Suess, E., editors, Proc.
ODP, Sci. Results, volume 146 (pt. 1), pages 175–186, College Station, TX.
Ocean Drilling Program.
Katz, M. E., Pak, D. E., Dickens, G. R., and Miller, K. G. (1999). The source and
fate of massive carbon input during the Latest Paleocene Thermal Maximum.
Science, 286:1531–1533.
Kennett, J. P., Cannariato, G., Hendy, I. L., and Behl, R. J. (2003). Methane Hy-
drates in Quaternary Climate Change: The Clathrate Gun Hypothesis. Am. Geo-
phys. Union, Washington DC.
Kimura, G., Silver, E. A., and Blum, et al., P., editors (1997). Proc. ODP, Init. Repts.,
volume 170, College Station, TX. Ocean Drilling Program.
Kleinberg, R. L., Flaum, C., Griffin, D. D., Brewer, P. G., Malby, G. E., Peltzer,
E. T., and Yesinowski, J. P. (2003). Deep sea NMR: Methane hydrate growth
habit in porous media and its relationship to hydraulic permeability, deposit
accumulation, and submarine slope stability. J. Geophys. Res., 108, 2508,
doi:10.1029/2003JB002389.
Kraemer, M., Owen, R. M., and Dickens, G. R. (2000). Lithology of the upper
gas hydrate zone, Blake Outer Ridge: A link between diatoms, porosity, and
gas hydrate. In Paull, C. K., Matsumoto, R., Wallace, P. J., and Dillon, W. P.,
editors, Proc. ODP, Sci. Results, volume 164, pages 229–236, College Station,
TX. Ocean Drilling Program.
Lee, M. W. (2000). Gas hydrates amount estimated form acoustic logs at the Blake
Ridge, Sites 994, 995, and 997. In Paull, C. K., Matsumoto, R., Wallace, P. J.,
and Dillon, W. P., editors, Proc. ODP, Sci. Results, volume 164, pages 193–198,
College Station, TX. Ocean Drilling Program.
Liu, X. and Flemings, P. B. (2006). Passing gas through the hydrate stability zone
at southern Hydrate Ridge, offshore Oregon. Earth Planet. Sci. Lett., 67:3403–
3421.
Luff, R. and Wallman, K. (2003). Fluid flow, methane fluxes, carbonate precipita-
tion and biogeochemical turnover in gas hydrate-bearing sediments at Hydrate
Ridge, Cascadia Margin: Numerical modeling and mass balances. Geochim.
Cosmochim. Acta, 67:3403–3421.
Luo, X., Vasseur, G., Pouya, A., Lamoureux-Var, V., and Poliakov, A. (1998).
Elastoplastic deformation of porous media applied to the modeling of compaction
at basin scale. Marine and Petr. Geol., 15:145–162.
MacKay, M. E., Jarrard, R. D., Westbrook, G. K., Hyndman, R. D., and the Ship-
board Scientific Party of Ocean Drilling Program Leg 146 (1994). Origin of
bottom-simulating reflectors: Geophysical evidence from the Cascadia accre-
tionary prism. Geology, 22:459–462.
Marshall, D. R., Saito, S., and Kobayashi, R. (1964). Hydrates at high pres-
sure, part i: Methane-water, argon-water, and nitrogen-water system. AICHE
J., 10:202–205.
Milkov, A. V., Claypool, G. E., Lee, Y. J., and R., S. (2005). Gas hydrate systems at
Hydrate Ridge offshore Oregon inferred from molecular and isotopic properties
of hydrate-bound and void gases. Geochim. Cosmochim. Acta, 69(4):1007–
1026.
Milkov, A. V., Dickens, G. R., Claypool, G. E., Lee, Y. J., Borowski, W. S., Torres,
M. E., Xu, W., Tomaru, H., Tréhu, A. M., and Schultheiss, P. (2004). Co-existence
of gas hydrate, free gas, and brine within the regional gas hydrate stability zone
at Hydrate Ridge (Oregon margin): evidence from prolonged degassing of a
pressureized core. Earth Planet. Sci. Lett., 222:829–843.
Milkov, A. V. and Sassen, R. (2000). Thickness of the gas hydrate stability zone,
Gulf of Mexico continental slope. Mar. Pet. Geol., 17:981–991.
Moridis, G. J. and Collett, T. S., editors (2003). Strategies for gas production from
hydrate accumulations under various geological and reservoir conditions. Pro-
ceedings, TOUGH Symposium, Lawrence Berkeley National Laboratory, Berke-
ley, California.
Niewöhner, C., Hensen, C., Kasten, S., Zabel, M., and Schulz, H. D. (1998). Deep
sulfate reduction completely mediated by anaerobic methane oxidation in sedi-
ments of the upwelling area off Namibia. Geochim. Cosmochim. Acta, 62:455–
464.
Paull, C. K., Matsumoto, R., Wallace, P. J., and Dillon, W. P., editors (2000). Proc.
ODP Sci. Results, volume 164, College Station, TX. Ocean Drilling Program.
Paull, C. K., Matsumoto, R., and Wallace, et al., P. J., editors (1996). Proceedings
of the Ocean Drilling Program, Initial Reports, volume 164, College Station, TX.
Ocean Drill. Program.
Paull, C. K., Ussler, W. I., and Borowski, W. S. (1994). Natural Gas Hydrates,
chapter Sources of biogenic methane to form marine gas hydrates, pages 392–
409. Ann. N.Y. Acad. Sci., 715.
253
Paull, C. K., Ussler III, W., Lorenson, T., Winters, W., and Dougherty, J. (2005).
Geochemical constraints on the distribution of gas hydrates in the Gulf of Mexico.
Geo. Mar. Lett., 25:273–280.
Piessens, R., de Doncker, E., Uberhuber, C. W., and Kahaner, D. K. (1983). Quad-
pack: A subroutine package for automatic integration. Springer-Verlag, New
York.
Riedel, M., Collett, T. S., Malone, M. J., and the Expedition 311 Scientists, editors
(2006). Proc. IODP, volume 311, Washington, DC. Integrated Ocean Drilling
Program Management International Inc.
Selim, M. S. and Sloan, E. D. (1989). Heat and mass-transfer during the dissocia-
tion of hydrates in porous-media. AICHE J., 35(6):1049–1052.
Seo, Y., Lee, H., and Ryu, B.-J. (2002). Hydration number and two-phase equilibria
of CH4 hydrate in the deep ocean sediments. Geophys. Res. Lett., 29, 1244,
doi:10.1029/2001GL014226.
Sloan, E. D., J. and Koh, C. (2007). Clathrate hydrates of natural gases. CRC
Press, third edition.
254
Snyder, G. T., Hiruta, A., Matsumoto, R., Dickens, G. R., Tomaru, H., Takeuchi, R.,
Komatsubara, J., Ishida, Y., and Yu, H. (2007). Pore water profiles and authigenic
mineralization in shallow marine sediments above the methane-charged system
on Umitaka Spur, Japan Sea. Deep-Sea Res. II, in press.
Subramanian, S., Ballard, A. L., Kini, R. A., Dec, S. F., and Sloan, E. D. (2000).
Structural transitions in methane plus ethane gas hydrates - part i: upper transi-
tion point and applications. Chem. Eng. Sci., 55(23):5763–5771.
Suess, E. and von Heune, et al., R., editors (1988). Proceedings of the Ocean
Drilling Program, Initial Reports, volume 112, College Station, TX. Ocean Drilling
Program.
Sultan, N., Cochonat, P., Foucher, J.-P., and Mienert, J. (2004). Effect of gas
hydrates melting on seafloor slope instability. Mar. Geol., 213:379–401.
Sun, X., Nanchary, N., and Mohanty, K. K. (2005). 1-D modeling of hydrate de-
pressurization in porous media. Transp. Porous Med., 58:315–338.
ten Haven, H. L., Littke, R., Rullkötter, J., Stein, R., and Welte, D. H. (1990). Ac-
cumulation rates and composition of organic matter in late Cenozoic sediments
underlying the active upwelling area off Peru. In Suess, E. and von Heune,
et al., R., editors, Proc. ODP, Sci. Results, volume 112, pages 591–606, College
Station, TX. Ocean Drilling Program.
Tishchenko, P., Hensen, C., Wallman, K., and Wong, C. S. (2005). Calculation of
the stability and solubility of methane hydrate in seawater. Chem. Geol., 219:37–
52.
Torres, M. E., Wallman, K., Tréhu, A. M., Bohrmann, G., Borowski, W. S., and
Tomaru, H. (2004). Gas hydrate growth, methane transport, and chloride en-
richment at the southern summit of Hydrate Ridge, Cascadia margin off Oregon.
Earth Planet. Sci. Lett., 226:225–241.
Tréhu, A. M., Bohrmann, G., Rack, F., and Torres, M. E., editors (2003). Proceed-
ings of the Ocean Drilling Program, Initial Reports, volume 204, College Station,
TX. Ocean Drill. Program.
255
Tréhu, A. M., Long, P. E., Torres, M. E., Bohrmann, G., Rack, F. R., Collett, T. S.,
Goldberg, D. S., Milkov, A. V., Riedel, M., Schultheiss, P., Bangs, N. L., Barr,
S. R., Borowski, W. S., Claypool, G. E., Delwiche, M. E., Dickens, G. R., Gra-
cia, E., Guerin, G., Holland, M., Johnson, J. E., Lee, Y. J., Liu, C. S., Su, X.,
Teichert, B., Tomaru, H., Vanneste, M., Watanabe, M., and Weinberger, J. L.
(2004). Three-dimensional distribution of gas hydrate beneath southern Hydrate
Ridge: Constraints from ODP Leg 204. Earth Planet. Sci. Lett., 222:845–862.
Ussler, W. and Paull, C. K. (2001). Ion exclusion associated with marine gas hy-
drate deposits. In Paull, C. K. and Dillon, W. P., editors, Natural Gas Hydrates:
Occurrence, Distribution, and Detection, volume 124 of Geophys. Monogr. Ser.,
pages 41–51. AGU, Washington, D. C.
van der Waals, J. H. and Platteeuw, J. C. (1959). Clathrate solutions. Adv. Chem.
Phys., 2:1–57.
Wallman, K., Aloisi, G., Haeckel, M., Obzhirov, A., Pavlova, G., and Tishchenko, P.
(2006). Kinetics of organic matter degradation, microbial methane generation,
and gas hydrate formation in anoxic marine sediments. Geochim. Cosmochim.
Acta, 70:3905–3927.
Wang, K., Hyndman, R. D., and Davis, E. E. (1993). Thermal effects of sedi-
ment thickening and fluid expulsion in accretionary prisms - model and parame-
ter analysis. J. Geophys. Res., 98(6):9975–9984.
Westbrook, G. K., Carson, B., and Musgrave et al., R. J., editors (1994). Proc.
ODP, Initial Reports, volume 146 (Pt. 1), College Statiom, TX. Ocean Drilling
Program.
Whiticar, M. J., Faber, E., and Schoell, M. (1986). Biogenic methane formation in
marine and freshwater environments: CO2 reduction vs. acetate fermentation -
Isotope evidence. Geochim. Cosmochim. Acta, 50:693–709.
Xu, W., Lowell, R. P., and Peltzer, E. T. (2001). Effect of seafloor temperature
and pressure variations on methane flux from a gas hydrate layer: Compari-
son between current and late Paleocene climate conditions. J. Geophys. Res.,
106:26413–26423.
256
Xu, W. and Ruppel, C. (1999). Predicting the occurrence, distribution, and evo-
lution of methane gas hydrate in porous marine sediments. J. Geophys. Res.,
104:5081–5096.
Yang, S. O., Cho, S. H., Lee, H., and Lee, C. S. (2001). Measurement and predic-
tion of phase equilibria for water + methane in hydrate forming conditions. Fluid
Phase Equil., 185:53–63.
Yuan, T., Hyndman, R. D., Spence, G. D., and Desmons, B. (1996). Seismic
velocity increase and deep-sea gas hydrate concentration above a bottom-
simulating reflector on the northern Cascadia continental slope. J. Geophys.
Res., 101:13655–13671.
Appendix A
Appendix
The net fluid flux in the system (Uf,tot ) results from the combination of fluid flux due
fluid flux (Uf,ext ) (Davie and Buffett, 2003b; Bhatnagar et al., 2007):
where Uf,sed can be related to the sedimentation rate and porosities as (Davie and
Buffett, 2003b):
1 − φ0
Uf,sed = Ṡφ∞ (A.2)
1 − φ∞
In terms of Peclet numbers (8.4), the sum Uf,tot can be written as:
Ds ∂c̃ls FSO4 Lt
(Pe1 + Pe2 ) c̃ls − φ = (A.4)
Dm ∂ z̃ ρf c0SO4 Dm
258
We now model porosity loss by relating it to effective stress and assuming hy-
between the reduced porosity and normalized depth (Bhatnagar et al., 2007):
η
φ̃ = (A.5)
η + (1 − η)ez̃
where φ̃ and η are reduced porosities defined in terms of the maximum (φ0 ) and
φ − φ∞ φ0 − φ∞
φ̃ = , η= (A.6)
1 − φ∞ 1 − φ∞
à ! à !
1+γ 1 + γ φ̃ Ds ∂c̃ls
(Pe1 + Pe2 ) c̃ls − = (A.7)
γ γ Dm ∂ z̃
à !
1 FSO4 Lt
, 0 < z̃ < L̃s
1 − φ∞ ρf c0SO4 Dm
The steady-state water mass balance below the SMT can be written as:
" #
∂ Us
Uf clw ρf + φSh chw ρh = 0 , Ls < z < Lt (A.8)
∂z 1−φ
where clw and chw are the water mass fractions in pore water and hydrate phase,
respectively. Equation (A.8) can also be written in terms of the water flux (FH2 O )
as:
Us
Uf clw ρf + φSh chw ρh = FH2 O (A.9)
1−φ
= (Uf,sed + Uf,ext )ρf
Due to low methane solubility in water, we assume the mass fraction of water in
aqueous phase to be unity. This gives us an expression for the water flux:
Us ρh
Uf = (Uf,sed + Uf,ext ) − φSh chw (A.10)
1−φ ρf
Substituting this expression for fluid flux into equation (9.25), we get:
" #
Us ρh l
(Uf,sed + Uf,ext ) − φSh chw c + (A.11)
1−φ ρf m
Us ρh ∂cl FCH4
φSh chm − φ(1 − Sh )Dm m =
1−φ ρf ∂z ρf
260
Similar to the sulfate mass balance, we multiply the above equation by (Lt /Dm )
" #
Pe1 Ũs
(Pe1 + Pe2 ) − φSh chw ρ̃h c̃lm + (A.12)
1−φ
Pe1 Ũs ∂c̃l FCH4 Lt
φSh c̃hm ρ̃h − φ(1 − Sh ) m =
1−φ ∂ z̃ ρf clm,eqb Dm
where different scaled variables were defined in equation (8.26). Equation (A.13)
Pe1 Ũs ³ ´
(Pe1 + Pe2 ) c̃lm + φSh ρ̃h c̃hm − chw c̃lm − (A.13)
1−φ
∂c̃l FCH4 Lt
φ(1 − Sh ) m =
∂ z̃ ρf clm,eqb Dm
The methane concentration profile in the absence of gas hydrate can be obtained
where D represents the bottom of the domain. These boundary conditions lead to
h ³ ´i
Q
1 − exp D̃s
g[z̃] − g[L̃s ]
c̃m (z̃) = c̃m,ext h ³ ´i , L̃s < z̃ < D (A.16)
1− exp D̃Qs g[D] − g[L̃s ]
If the depth of the external boundary condition (D) is sufficiently greater than unity,
zero, simplifying the methane concentration profile and the methane flux as follows:
à " #!
Q ³ ´
c̃m (z̃) = c̃m,ext 1 − exp g[z̃] − g[L̃s ] , L̃s < z̃ < D and D À 1 (A.17)
D̃s
Equating this methane flux to the sulfate flux at the SMT using equations (8.15)
Q
(−m)Qc̃m,ext = h
Q
³ ´i (A.19)
1 − exp D̃s
g[0] − g[L̃s ]
The above equation can be used to obtain the modified flux, Q, as a function of L̃s :
³ ´
1+mc̃m,ext
D̃s ln mc̃m,ext
Q= (A.20)
g[0] − g[L̃s ]
Finally, substituting the above relationship and equation (A.17) into the tangency
conditions (8.45) and (8.46) gives the coupled set of equations (8.47) and (8.48).
263
The accumulation terms in the mass balance equations (10.26) or (10.37) contain
products of primary variables such as porosity (φ), concentration (c̃lm ) and satura-
an example, we use the following expansion for the product (φSw c̃lm ) to conserve
mass:
where ∆t denotes the time change operator and superscripts n+1 and n represent
The above equation can be rewritten in terms of the time change operator as:
lithostatic stress and pore pressure. Hence, ∆t can be written as a function of pore
264
∂φ
∆t (φSw clm ) ≡ φn+1 ∆t (Sw clm ) + Swn cl,n
m ∆t pw (A.25)
∂pw
à !à !
φ − φ∞ σvn+1 − σvn
∆t (φSw clm ) ≡ φn+1 ∆t (Sw clm ) + Swn cl,n
m 1 − n+1 ∆t pw (A.26)
σφ pw − pnw
The lithostatic stress values are evaluated by integrating the sediment bulk density
gradient, equation (10.16), and are updated during the iterations. Similar expan-
The coupled equations (10.26), (10.31), (10.34) and (10.37) are solved numerically
using a fully implicit formulation, with the primary variables being p̃w , ṽs , α̃ and one
of the following three (c̃lm ,Sh , Sg ). Choice between these last three primary variables
given timestep. Schemes for switching among different primary variables in similar
problems have been proposed in the literature (Falta et al., 1992; Liu and Flemings,
2007). In our model, for example, if the gridblock is undersaturated with respect
to methane, both hydrate and free gas saturations are zero and c̃lm becomes the
greater than the local solubility, c̃lm becomes fixed, whereas Sh or Sg become the
primary variable, depending on whether the gridblock is within the region of hydrate
or stability or not.
Three possible states are identified for switching between different primary vari-
ables (c̃lm ,Sh , Sg ) in the water and methane mass balances. Each gridblock is as-
signed a phase state, which can change in response to the local thermodynamic
Phase state 1: Pore water is undersaturated with respect to methane at the previ-
ous time step and remains undersaturated at the new time step. Thus, if c̃lm,sol is
266
methane solubility at any grid block (i, j), then for this phase state,
Phase state 2: Pore water is undersaturated with respect to methane at the previ-
ous time step but becomes saturated at the new time step.
l,n+1
primary variable and c̃m = c̃lm,sol
Phase state 3: Pore water is saturated with respect to methane at the previous
n+1
Phase (i, j) = 3 if c̃l,n l
m = c̃m,sol and Sh or Sgn+1 > 0 ⇒ Shn+1 or Sgn+1 remains the
l,n+1
primary variable and c̃m = c̃lm,sol
Hydrate or free gas saturation going below zero during any iteration implies the
disappearance of hydrate or free gas phases. This means that dissolved methane
concentration can vary instead of being constrained by the solubility curve. Con-
ABSTRACT
The effect of overpressure on gas hydrate and free gas distribution in marine sediments is studied
using a one-dimensional numerical model that couples sedimentation, fluid flow, and gas hydrate
formation. Natural gas hydrate systems are often characterized by high sedimentation rates and/or
low permeability sediments, which can lead to pore pressures higher than hydrostatic. To quantify
the relative importance of these two factors, we define a dimensionless sedimentation-compaction
group, N sc , that compares the absolute permeability of the sediments to the sedimentation rate.
Higher values of N sc mean higher permeability or low sedimentation rate implying hydrostatic
pore pressure. Conversely, lower values of N sc indicate pore pressures greater than hydrostatic.
Simulation results show that decreasing N sc not only increases pore pressure from hydrostatic
values, but also lowers the lithostatic stress gradient and gas hydrate saturation. This occurs
because excess pore pressures result in smaller effective stress, causing high porosity and lower
bulk density of the sediment. This leads to higher sediment velocity through the stability zone,
thereby reducing the mass accumulation of methane and gas hydrate in the pore space. Effect of
overpressure on depth of the gas hydrate stability zone is also studied.
∗
Corresponding author: Phone: +1 713 348 5416 Fax +1 713 348 5478 E-mail: gjh@rice.edu
pj Pressure of phase j increase in overburden is faster than rate of pore
water expulsion [3].
Sj Saturation of phase j
S Sedimentation rate at the seafloor Overpressure impacts the behavior of gas hydrate
t Time systems in several ways. For example, the
maximum thickness of the free gas layer below the
U f , sed Fluid flux due to sedimentation
base of the gas hydrate stability zone (GHSZ)
vj Velocity of phase j depends on the magnitude of overpressure [2,6,7].
z Depth below seafloor The length of a free gas column sealed by
α Organic carbon content overlying gas hydrate is regulated by the
difference between pore water pressure and
α0 Organic carbon content at seafloor
lithostatic stress. Thus, higher pore water pressures
β Normalized organic content at seafloor imply that relatively short connected gas columns
γ,η Reduced porosity parameters can develop before fracturing or shear failure
λ Methanogenesis reaction rate occurs, thereby causing a sudden release of free
gas [2,6,7]. Conversely, relatively long connected
μj Viscosity of phase j
gas columns can form when water overpressure is
ρj Density of phase j zero.
σv Vertical effective stress
Overpressure also affects sediment and gas
σφ Characteristic stress for compaction hydrate velocity through the GHSZ. At hydrostatic
φ Porosity pore pressures, sediments achieve maximum
compaction due to relatively high effective
φ0 Porosity at seafloor stresses acting on them. Development of
φ∞ Porosity at great depths overpressure reduces the effective stress acting on
the sediments, resulting in higher sediment
porosity, which leads to faster sediment velocity
Subscripts/superscripts:
through the GHSZ. In a gas hydrate system
g Gas phase
dominated by in-situ biogenic methane supply, this
h Hydrate phase increase in sediment velocity curtails the amount
l , w Water phase or component of organic carbon converted within the GHSZ.
m Methane component This occurs because the organic carbon is also
s Sediment phase progressively buried deeper with the sediment.
Additionally, increase in sediment velocity also
reduces the residence time of gas hydrate in the
INTRODUCTION GHSZ. These two mechanisms can cause
overpressure to result in relatively lower gas
Gas hydrate systems can be characterized by hydrate and free gas saturations at steady-state.
overpressure, i.e. pore water pressures higher than
hydrostatic. This is particularly evident at settings Increase in pore pressure can also change the
dominated by low permeability silts/clays, e.g., thermodynamic stability regime and extend the
Blake Ridge [1,2]. Over geologic timescales, base of the GHSZ to greater depths below the
continuous sedimentation causes increase in the seafloor. However, we show through numerical
overburden, resulting in compaction of sediments simulations that this increase in depth of the
[3,4,5]. Overpressure can develop in such systems GHSZ due to overpressure is relatively small, even
if pore water cannot be expelled from the pore when pore pressures become close to lithostatic
space fast enough and, instead, starts to support limits.
some of the overburden. Since permeability
controls this rate of pore water expulsion, We have previously developed generalized
sediments with low permeability can develop dimensionless numerical models to study gas
overpressure [4,5]. Alternatively, overpressure can hydrate and free gas distribution in marine
also develop in sediments with relatively high sediments [8,9]. However, in the previous work,
permeability if the sedimentation rate is fast, i.e. pressure was assumed to be hydrostatic. In this
paper, we extend the previous dimensionless Organic Balance:
models to explicitly incorporate water pressure
through the use of Darcy’s law in a compacting ∂ ∂
medium. This allows us to model overpressure ⎡⎣(1 − φ ) ρ s ⎤⎦ α + ⎡⎣(1 − φ ) ρ s vsα ⎤⎦ =
∂t ∂z (4)
development in compacting sediments and study
its effect on gas hydrate/free gas saturation. − ρ s λ (1 − φ ) α
The water, methane, sediment and organic mass Water relative permeability in the presence of gas
balances are written as: hydrate is (assuming pore-filling structure) [12]:
Water Balance:
2 (1 − Sh )
2
krw = 1 − S +
2
(7)
ln ( Sh )
h
∂
⎡⎣φ S w cwl ρ w + φ Sh cwh ρ h ⎤⎦ +
∂t
(1) Water relative permeability in the presence of free
∂
⎡⎣φ S wcwl ρ w vw + φ Sh cwh ρ h vs ⎤⎦ = 0 gas is [10]:
∂z
S w − S wr
( )
4
Methane Balance: krw = krw
0
S w* , where S w* = (8)
1 − S wr
∂
⎡⎣φ S wcml ρ w + φ Sh cmh ρ h + φ S g cmg ρ g ⎤⎦ + Porosity-effective stress relationship is defined as
∂t [13]:
∂
⎡⎣φ S w cml ρ w vw + φ S h cmh ρ h vs + φ S g cmg ρ g vs ⎤⎦ (2) σ v − pw
∂z −
φ = φ∞ + (φ0 − φ∞ ) e
σφ
(9)
∂ ⎡ ∂c l ⎤ M CH 4
= ⎢φ S w Dm ρ w m ⎥ + ρ s λ (1 − φ ) α
∂z ⎣ ∂z ⎦ M org Lithostatic stress gradient can be written as a
function of densities and porosity as:
Sediment Balance:
∂σ v
∂ ∂ = ⎡(1 − φ ) ρ s + φρ w ⎤⎦ g
∂z ⎣
(10)
⎡⎣(1 − φ ) ρ s ⎤⎦ + ⎡⎣(1 − φ ) ρ s vs ⎤⎦ = 0 (3)
∂t ∂z
Normalized variables and dimensionless groups Lithostatic stress ( σ v ), water pressure and gas
pressure are normalized by hydrostatic water
The above equations are now written in pressure at the base of the GHSZ:
dimensionless form. Reduced porosities are
defined as:
σv pw pg
σ v = , p w = , p g = (16)
φ − φ∞ φ − φ∞ 1 − φ∞ ρ w gLt ρ w gLt ρ w gLt
φ = ,η= 0 , γ = (11)
1 − φ∞ 1 − φ∞ φ∞
Dimensionless depth and time are defined as:
The dimensionless Peclet number and Damkohler
z t
number are defined as: z = , t = 2 (17)
Lt Lt / Dm
U f , sed Lt λ L2t
Pe1 = , Da = (12) All phase densities are normalized by water
Dm Dm
density ( ρ i = ρi / ρ w ). Sediment velocity is
We also define a dimensionless group normalized by the sedimentation rate at the
corresponding to the ratio of the absolute sediment seafloor:
permeability and the sedimentation rate at the
seafloor: v
vs = s (18)
S
k0 ρ w g
N sc = (13)
μ w S Finally, organic carbon content and initial carbon
content at the seafloor are scaled as:
Large values of N sc correspond to high sediment
α α
permeability and/or low sedimentation rate, α = , β= l 0 (19)
implying hydrostatic pressures. Conversely, low α0 cm ,eqb
values of N sc imply low permeability and/or high
The resulting dimensionless mass balances and
sedimentation rate, thereby causing pore pressures
constitutive relationships are given in the
higher than hydrostatic. Similar dimensionless
appendix.
groups have been used to model non-hydrostatic
compaction in sedimentary basins [14,15].
NUMERICAL SOLUTION
The ratio of compaction depth to the base of
The coupled dimensionless equations are solved
GHSZ is defined by the dimensionless group,
numerically using a fully implicit finite difference
N tφ :
formulation, with the primary variables being p w ,
vs , α and one of the following three
Lφ σ φ / ( ρw g )
N tφ = = (14) ( cml , S h , S g ). Choice between these last three
Lt Lt
primary variables is made according to the local
thermodynamic conditions at any gridblock at any
The normalized methane concentrations are
given time-step. All four mass balances are cast in
defined as:
residual form and the Newton-Raphson method is
used to iterate on them to converge to the solution.
cml cmh cmg
cml = , cmh = , cmg = (15)
cml ,eqb cml ,eqb cml ,eqb RESULTS
But even when pore pressures are close to However, it should be noted that hydrate and free
lithostatic, the downward shift in the base of the gas saturation profiles do not give a complete
GHSZ is very small in the normalized form. When picture of their amounts, because each value of
the normalized vertical depth scale (Figure 3) is N sc results in a different porosity profile. Thus,
converted back to the physical scale by although hydrate and free gas saturation within the
multiplying with Lt , the depth to the base of pore space decrease on lowering N sc , the
GHSZ, this increase in the thickness of the GHSZ corresponding increase in porosity might lead to
net higher accumulation of hydrate or free gas columns may result for settings characterized by
within the sediment volume. To test this scenario, high N sc . In contrast, only short gas columns can
we plot the product of porosity and hydrate/free
form when N sc is low before sediment
gas saturation ( φ S j ) to get the volume fraction of
fracture/failure occurs and vents the free gas into
hydrate and free gas within the sediment. These the ocean. Thus, from an exploration standpoint,
profiles are plotted in Figure 5 and show that the geologic sites characterized by high permeability
net amount of gas hydrate or free gas saturation
and low sedimentation rates (i.e., high N sc ) might
within the sediment also decreases on lowering
N sc . However, multiplying by porosity does be most suitable for targeting the free gas sealed
by a hydrate layer.
reduce the magnitude of change observed between
different cases. For example, peak hydrate CONCLUSIONS
saturation at the base of GHSZ decreases from
about 6% to 1%, a factor of 6 change, on lowering A dimensionless numerical model for non-
N sc from 10000 to 2. In contrast, peak change in hydrostatic pressure compaction is developed to
sediment volume fraction of hydrate goes from study the effect of overpressure on gas hydrate and
about 2.2% to 0.7%, a factor of 3 change, for the free gas saturations. Non-dimensionalization of the
same decrease in N sc . equations lead to a sedimentation-compaction
group, N sc , defined as the ratio of sediment
permeability to sedimentation rate. Simulations
show that relatively high values of N sc (about
104) lead to systems close to hydrostatic pore
pressure, while relatively low values of N sc
(about unity) lead to significant overpressure in the
system. Overpressure development impacts this
gas hydrate system by lowering effective stresses
on the sediment, causing higher porosities. Higher
sediment velocities achieved due to overpressure
and high porosities ultimately lead to lesser
organic carbon decay, resulting in lower hydrate
and free gas saturations for our set of boundary
Figure 5: Effect of overpressure, characterized conditions, i.e., fixed seafloor depth and constant
through N sc , on steady-state gas hydrate (solid geotherm.
curves) and free gas (dashed curves) sediment
volume fraction.
REFERENCES
This validates our hypothesis that overpressure
does lower the net amount of methane that [1] Winters WJ. Stress history and geotechnical
accumulates in either hydrate or free gas phase. In properties of sediment from the Cape Fear Diapir,
other words, the decrease in hydrate and free gas Blake Ridge Diapir, and Blake Ridge. In: Paull
saturation with increasing overpressure is not only CK, Matsumoto R, Wallace PJ, Dillon WP,
a result of increased porosities. The effect of editors. Proceedings of the Ocean Drilling
increased sediment and fluid velocities and lower Program, Scientific Results, volume 164: College
organic carbon decay within the GHSZ has a much Station, Texas Ocean Drilling Program, 2000. p.
more significant impact on net gas hydrate and 421-429.
free gas accumulation. [2] Flemings PB, Liu X, Winters WJ. Critical
pressure and multiphase flow in Blake Ridge gas
Effect on free gas column thickness hydrates. Geology 2003; 31:1057-1060.
[3] Gibson RE. The progress of consolidation in a
Based on pressure profiles shown in Figures 1 and clay layer increasing in thickness with time.
2, it can be argued that deep connected free gas Geotechnique 1958; 8:171-182.
[4] Wangen M. Pressure and temperature [16] Paull CK, Matsumoto R, Wallace PJ, Dillon
evolution in sedimentary basins. Geophysical WP, editors. Proceedings of the Ocean Drilling
Journal International 1992; 110:601-613. Program, Scientific Results, volume 164: College
[5] Dugan B, Flemings PB. Overpressure and fluid Station, Texas Ocean Drilling Program, 2000. p.
flow in the New Jersey continental slope: 421-429.
implications for slope failure and cold seeps.
Science 2000; 289:288-291. APPENDIX
[6] Hornbach MJ, Saffer, DM, Holbrook, WS.
Critically pressured free-gas reservoirs below gas- Dimensionless mass balances
hydrate provinces. Nature 2004; 427:142-144. The scaling schemes defined in the main text lead
[7] Liu XL, Flemings PB. Dynamic multiphase to the following form of the four mass balances,
flow model of hydrate formation in marine initial conditions (I.C.) and boundary conditions
sediments. Journal of Geophysical Research 2007; (B.C.).
112:B03101, doi:10.1029/2005JB004227.
[8] Bhatnagar G, Chapman WG, Dickens GR, Water Balance:
Dugan B, Hirasaki GJ. Generalization of gas
hydrate distribution and saturation in marine ∂ ⎡1 + γφ ⎤ ⎛ 1+ γ ⎞
sediments by scaling of thermodynamic and ⎢
∂t ⎣ γ
( )
S wcwl + Sh cwh ρ h ⎥ + Pe1 ⎜ ⎟
transport processes. American Journal of Science ⎦ ⎝ 1 −η ⎠
2007; 307:861-900. ⎡1 + γφ 1 + γ ⎛ 1 + γφ ⎞ ⎤
8
[9] Bhatnagar G, Chapman WG, Dickens GR, ⎢ S w cw vs − N sc
l
⎜ ⎟ ⎥
Dugan B, Hirasaki GJ. Sulfate-methane transition ∂ ⎢ γ γ ⎝ 1 + γη ⎠ ⎥
as a proxy for average methane hydrate saturation ∂z ⎢ ⎛ ∂p ⎥=0
⎢k ⎞ + γφ
Sh cwh ρ h vs ⎥
1
in marine sediments. Geophysical Research w
− 1⎟ cwl +
Letters 2008; 35:L03611, ⎢⎣ rw ⎜⎝ ∂z ⎠ γ ⎥⎦
doi:10.1029/2007GL032500. (A1)
[10] Bear J. Dynamic of fluids in porous media.
New York: Dover Publications Inc., 1988. ρ w gD0 + ρ w gz D0
[11] Smith JE. The dynamics of shale compaction I.C.: p w ( z, 0 ) = = + z (A2)
and evolution of pore-fluid pressures. ρ w gLt Lt
Mathematical Geology 1971; 3:239-263.
[12] Kleinberg RL, Flaum C, Griffin DD, Brewer D0
PG, Malby GE, Peltzer ET, Yesinowski JP. Deep B.C. (1): p w ( 0, t ) = (A3)
Lt
sea NMR: Methane hydrate growth habitat in
∂p w
porous media and its relationship to hydraulic
permeability, deposit accumulation, and
B.C. (2):
∂z
( D, t ) = 1 (A4)
submarine slope stability. Journal of Geophysical
Research 2003; 108(B10):2508, where D0 is seafloor depth and D is the bottom
doi:10.1029/2003JB002389.
[13] Rubey WW, Hubbert MK. Role of fluid of the domain.
pressure in mechanics of over-thrust faulting; ii.
Overthrust belt in geosynclinal area of western
Wyoming in light of fluid pressure hypothesis.
Geological Society of America Bulletin 1959;
70:167-206.
[14] Yang XS, Fowler AC. Fast and slow
compaction in sedimentary basins. SIAM Journal
on Applied Mathematics 1998; 59(1):365-385.
[15] Gutierrez M, Wangen M. Modeling of
compaction and overpressuring in sedimentary
basins. Marine and Petroleum Geology 2005;
22(3):351-363.
Methane Balance: B.C. (1): α ( 0, t ) = 1 (A12)
∂ ⎡1 + γφ ⎤
⎢
∂t ⎣ γ
( )
S wcml + S h cmh ρ h + S g cmg ρ g ⎥ + Pe1 Initial porosity profile
⎦ Reduced porosity ( φ ) is related to the
⎡1 + γφ ( σ v )
1 + γ ⎛ 1 + γφ ⎞ ⎤
8 dimensionless lithostatic stress and
⎢ S wcml vs − N sc ⎜ ⎟ ⎥ dimensionless pore pressure ( p w ):
⎢ γ γ ⎝ 1 + γη ⎠ ⎥
⎢ ⎥
⎢ k ⎛ ∂p w − 1⎞ c l + 1 + γφ S c h ρ v ⎥ ⎡ σ v − p w ⎤
rw ⎜ ⎟ m
⎛ 1 + γ ⎞ ∂ ⎢ ⎝ ∂z ⎠ γ h m h s
⎥ φ = η exp ⎢ − ⎥ (A13)
⎜ ⎟ ⎢ ⎥ ⎣⎢ N tφ ⎦⎥
⎝ 1 − η ⎠ ∂z ⎢ 1 + γφ g 1+ γ ⎥
⎢ + γ S g cm ρ g vs − N sc γ ⎥ At hydrostatic pressure, the porosity profile can be
⎢ ⎥
⎢⎛ 1 + γφ ⎞8 ⎛ ∂p g ⎥ computed as an analytical expression to serve as
⎞μ
⎢⎜ ⎟ krg ⎜ − ρ g ⎟ w ρ g cml ⎥ an initial condition:
⎣⎢⎝ 1 + γη ⎠ ⎝ ∂z ⎠ μg ⎦⎥
η
∂ ⎡1 + γφ ∂cml ⎤ M CH 4 φ =
( )
(A14)
= ⎢ Sw ⎥+ ρ s Da 1 − φ αβ
⎡ γ z ( ρ s − 1) ⎤
∂z ⎣ γ ∂z ⎦ M org η + (1 − η ) exp ⎢ ⎥
(A5) ⎢⎣ N tφ (1 + γ ) ⎥⎦
Sediment Balance:
∂ ⎛ 1+ γ ⎞ ∂ ⎡
∂t
⎡1 − φ ⎤ + Pe1 ⎜
⎣ ⎦ − η ∂
(
)
⎟ ⎣ 1 − φ vs ⎤⎦ = 0
⎝ 1 ⎠ z
1 −η
I.C.: vs ( z, 0 ) = (A9)
1 − φ
Organic Balance:
∂ ⎡ ⎛ 1+ γ ⎞ ∂ ⎡
∂t ⎣ ( )
1 − φ α ⎤ + Pe1 ⎜
⎦ − η ∂
(
⎟ ⎣ 1 − φ vsα ⎤⎦ )
⎝ 1 ⎠ z
(
= − Da 1 − φ α )
I.C.: α ( z , 0 ) = 0 (A11)
The sulfate-methane transition as a proxy for average
methane hydrate saturation in marine sediments
G. Bhatnagar1, W. G. Chapman1, G. R. Dickens2, B. Dugan2, G. J. Hirasaki1*
1
Department of Chemical & Biomolecular Engineering, Rice University, Houston,
Texas
2
Department of Earth Science, Rice University, Houston, Texas
*
Corresponding author: gjh@rice.edu
Abstract.
We develop a relationship between the sulfate-methane transition (SMT)
and average gas hydrate saturation (AGHS) for systems dominated by methane
migration from deeper sources. The relationship is explained by a one-
dimensional numerical model that simulates gas hydrate accumulation in marine
sediments. Higher methane fluxes result in shallow SMT depths and high AGHS,
while lower methane fluxes result in deep SMTs and low AGHS. We also
generalize the variation between AGHS and scaled SMT depth, a procedure that
aids prediction of AGHS at different sites from observations of the SMT, such as
along Cascadia Margin.
1. Introduction
Gas hydrates can form in the pore space of sediment along continental
margins when methane and other low molecular weight gases combine with
water at appropriate pressure, temperature and salinity conditions [Kvenvolden,
1993]. These hydrates are components of dynamic systems in which methane
enters and leaves a gas hydrate stability zone (GHSZ) at variable rates [Dickens,
2003]. Based on the supply of methane, marine gas hydrate systems can be
distinguished into two end-members: in-situ systems where microbes generate
methane within the GHSZ [e.g., Claypool and Kvenvolden, 1983]; and deep-
source systems where rising fluids bring methane from depth [e.g., Hyndman and
Davis, 1992]. However, quantifying gas hydrate saturation in these systems
remains a challenge. In this paper, we develop a model that relates average gas
hydrate saturation (AGHS) to the depth of sulfate-methane transition (SMT) in
deep-source systems. This facilitates prediction of AGHS at sites where SMT
depth is known.
The SMT denotes a relatively thin zone near the seafloor where pore
water sulfate and methane are depleted to zero concentration (Figure 1). This
depletion occurs due to the anaerobic oxidation of methane reaction
(AOM: CH 4 + SO42− → HCO3− + HS − + H 2O ) [Borowski et al., 1999]. Although
microbes can also consume sulfate using solid organic carbon [Berner, 1980],
AOM can dominate overall sulfate depletion in sediments with gas hydrates and
1
modest methane fluxes [Borowski et al., 1996; Snyder et al., 2007]. Further,
since we focus only on deep-source gas hydrate systems (i.e., sites with low
organic carbon input), AOM becomes the only sulfate sink. The sulfate profile
and SMT depth in such deep-source systems should depend on methane flux
from below because of the simple 1:1 AOM reaction [Borowski et al., 1996;
Snyder et al., 2007]. Additionally, the thickness of the gas hydrate zone and gas
hydrate saturation are functions of upward methane flux [Davie and Buffett, 2003;
Bhatnagar et al., 2007]. Thus, SMT depth ( Ls , Figure 1) should relate to gas
hydrate saturation [Borowski et al., 1999].
To study this relationship between SMT depth and AGHS, we expand the
model of Bhatnagar et al. [2007] by including a sulfate balance for deep-source
systems (Figure 1). We show that, at steady-state conditions, the depth of the
SMT relates to net fluid flux in the system and to AGHS (volume fraction of pore
space) within the GHSZ. Compared to previous site-specific studies, our model
generalizes the relationship between SMT depth and AGHS at any gas hydrate
setting dominated by methane flux from depth.
2
Methane mass fraction in phase i ( cmi ) is scaled by methane solubility in
the liquid phase at the base of GHSZ ( cmw,eqb ), cwh is mass fraction of water in
hydrate phase ( cwh = 1 − cmh ), while sulfate mass fraction in pore water ( csw ) is
scaled by the seawater value ( csw,0 ), to get the corresponding normalized
variables:
cmi csw
cmi = for i ∈ {w, h, g} , csw = (2)
cmw,eqb csw,0
where U f ,ext is the upward fluid flux due to external sources and has negative
value (due to opposite direction to U f , sed ), and λ AOM is the second order rate
constant for AOM. Thus, Pe1 has positive value, while Pe2 becomes negative.
Since we focus on deep-source systems, results shown later are relevant for
cases where Pe2 > Pe1 . Importantly, Pe1 characterizes the ratio of compaction-
driven fluid flux to methane diffusion, while Pe2 represents the ratio of external
3
fluid flux to methane diffusion. The Damkohler number compares AOM rate to
methane diffusion. Finally, we complete the system by formulating the
dimensionless sulfate mass balance:
∂ ⎡1 + γφ ⎤ 1 + γ ∂ ⎡⎛ (1 + γφ ) ⎞ ⎤
⎢ S wcsw ⎥ + ⎢ ⎜ ( Pe1 + Pe2 ) − Pe1U s Sh cwh ρ h ⎟ csw ⎥ =
∂t ⎣ γ ⎦ γ ∂z ⎣⎢⎝ γ (1 − φ ) ⎠ ⎦⎥
(7)
∂ ⎡1 + γφ ⎛ Ds ⎞ ∂csw ⎤ ⎡1 + γφ ⎤ w w
⎢ Sw ⎜ ⎟ ⎥ − Da AOM ⎢ S w ⎥ cm cs
∂z ⎢⎣ γ ⎝ Dm ⎠ ∂z ⎥⎦ ⎣ γ ⎦
where Ds denotes sulfate diffusivity. The initial and boundary conditions for the
two mass balances are written as:
I.C.: csw ( z, 0) = cmw ( z, 0) = 0 (8)
∂csw cmw,ext
B.C. (2): ( D, t ) = 0 , cmw ( D, t ) = cmw,ext = (10)
∂z cmw,eqb
where cmw,ext is the methane concentration in the external flux, cmw,ext is the
normalized value, and D denotes the bottom of the model domain.
3. Results
Equations (1) and (7) are solved numerically to obtain steady-state profiles
for methane, gas hydrate saturation, and pore water sulfate concentration. For
results shown later, we assume seafloor temperature ( T0 ) to be 3°C, geothermal
gradient ( G ) to be 0.04°C/m, and pore water salinity representative of standard
seawater. Changing T0 or G results in methane solubility curves that are similar
in the normalized form [Bhatnagar et al., 2007], causing AGHS to be relatively
insensitive to changes in T0 or G . However, the normalized solubility curves are
more sensitive to seafloor depth. Thus, we use a seafloor depth of 1000 m for
results shown in Figures 2 and 3, whereas Figure 4 generalizes the relationship
between AGHS and SMT depth for multiple seafloor depths. Porosity at the
seafloor ( φ0 ) and at depth ( φ∞ ) are assumed to be 0.7 and 0.1, respectively.
Diffusivities Ds and Dm are taken to be 0.56×10-9 and 0.87×10-9 m2/s,
respectively [Iversen and Jørgensen, 1993], cmh is set to 0.134, seawater sulfate
concentration equals 28 mM, and ρ h and ρ f equal 930 and 1030 kg/m3,
respectively. At steady state, cmw,ext is not significant, provided it exceeds the
minimum required to form hydrate [Bhatnagar et al., 2007]. Consequently, we
assume in all simulations here that cmw,ext equals unity.
4
We first study the effect of Da AOM on steady-state profiles. For fixed Pe1
and Pe2 , decreasing Da AOM results in a thickening of the SMT zone (Figure 2a).
Higher Da AOM implies faster consumption of methane and sulfate compared to
diffusion, causing a relatively sharp SMT. The thickness of the SMT zone is
usually less than a few meters at most gas hydrate settings, so we use a large
value of Da AOM (108) in further simulations.
Concentration profiles simulated for three different sets of Pe1 and Pe2 ,
but with the sum Pe1 + Pe2 held constant at -10, are shown in Figure 2b. Overlap
of these profiles demonstrates that neither Pe1 nor Pe2 individually controls the
concentration profiles, but that their sum determines the concentrations and the
scaled SMT depth, Ls = Ls / Lt . This sum, Pe1 + Pe2 , represents the net fluid flux
through the system. Hydrate saturation profiles, however, depend on more than
the sum of the Peclet numbers (Figure 2c). The AGHS ( S h ) for each of the
three cases is about 0.2%, 0.5% and 2%, with the highest value corresponding to
the smallest Pe1 (0.1) and largest Pe2 (-10.1). Small Pe1 and large Pe2
correspond to low sedimentation rate and high methane flux, respectively,
resulting in higher AGHS. However, for all three cases, the product Pe1 S h is the
same. Thus, Figure 2c demonstrates that Pe1 S h , which characterizes the flux of
gas hydrate through the GHSZ, is controlled by the net fluid flux, Pe1 + Pe2
[Bhatnagar et al., 2007].
Increasing net methane flux from depth (i.e., raising the magnitude of
Pe1 + Pe2 ) results in a shallow scaled SMT depth (Figure 3a), as proposed by
Borowski et al. [1996, 1999]. Increasing Pe2 , with Pe1 held constant, increases
gas hydrate saturation (Figure 3b) due to higher methane input to the system.
Consequently, Pe1 S h also increases. Hence, the scaled depth to the SMT, Ls ,
and Pe1 S h both depend on the sum Pe1 + Pe2 . As a consequence, scaled SMT
depth and Pe1 S h become correlated. This correlation, shown in Figure 4 for
three seafloor depths, indicates that average gas hydrate flux, Pe1 S h , increases
as Ls decreases. Thus, AGHS can be estimated for any system dominated by
methane flux from depth if L and Pe are known (Figure 4).
s 1
5
controlled by methane supplied from depth [Riedel et al., 2006]. We now
summarize calculation of AGHS from site-specific data at Cascadia Margin Sites
889, U1325, U1326 and U1328 (Table 1):
• Use sedimentation rate ( S ) to calculate Pe1 from equations (5) and (6);
• Calculate the scaled SMT depth Ls using the dimensional depths Ls and
Lt ;
• For given seafloor depth and Ls , obtain gas hydrate flux Pe1 S h from
Figure 4;
• Divide this gas hydrate flux by Pe1 to yield AGHS, S h (Table 1)
At Site 889 (ODP Leg 146), pore water chloride profile indicates a peak
hydrate saturation close to 2% at the base of GHSZ, and AGHS <1% within the
GHSZ [Davie and Buffett, 2003]. This result agrees favorably with our simulation
that shows peak saturation of about 2.2 % at the base of GHSZ and AGHS of
0.4% across the entire GHSZ (Table 1). Hyndman et al. [1999] calculated gas
hydrate saturation between 25-30% of pore space in the 100 m interval above
the base of GHSZ at Site 889 using resistivity log data. Subsequent calculations
using a different set of Archie parameters revise this estimate to 5-10% in that
100 m interval [Riedel at al., 2006]. Further, Ussler and Paull [2001] show that a
smoothly decreasing chlorinity profile at Site 889 yields hydrate saturation of 2-
5% within discrete layers. Although several parameter uncertainties confront
such estimates [Egeberg and Dickens, 1999; Riedel et al., 2006], AGHS
predicted using our SMT model concurs with the lower estimates at Site 889.
For the IODP Expedition 311 sites, drilled along the northern Cascadia
Margin, we compare our predictions with AGHS computed from chloride
anomalies and resistivity log data (Table 1). AGHS is calculated from chloride
data by assuming a background in situ chloride profile and attributing the relative
pore water freshening to gas hydrate dissociation [e.g., Egeberg and Dickens,
1999]. AGHS is obtained from resistivity data using the Archie equation and
parameters given in Riedel et al. [2006]. AGHS at Sites U1325 and U1326
estimated from resistivity and chlorinity are similar and our predictions based on
SMT depth are close to these estimates (Table 1). At Site U1328, our predicted
AGHS is distinctly lower than resistivity-based estimate (Table 1). Site U1328 is a
cold vent characterized by focused fluid and gas flow that causes high gas
hydrate saturations close to the seafloor [Riedel et al., 2006]. Such local
heterogeneities that might enhance methane flux from depth are not included in
our simple 1-D model, thereby causing greater deviation between predicted and
estimated AGHS.
Overall, we get good first order agreement between AGHS derived from
chloride anomalies/resistivity logs and those predicted using our model, although
our simulations consistently show lower AGHS at these sites. A possible
explanation for this general deviation is that interpretations of resistivity logs
6
depend on knowledge of formation water resistivity and three empirical constants,
which are hard to constrain in clay-rich sediments. Additionally, our simulations
(Figures 2 and 3) and previous models [e.g., Davie and Buffett, 2003] predict gas
hydrate to first occur well below the seafloor. In contrast, log-based results often
predict hydrate starting immediately below the seafloor, causing AGHS to be
higher than that predicted from simple transport models. Apart from the small
deviations between model and chloride/resistivity log predictions, our model
captures the trend in the lateral variation of AGHS correctly and likely provides a
lower bound on AGHS. Hence, our model and generalized results (Figure 4)
provide a simple and fast technique to constrain AGHS in deep-source gas
hydrate systems.
5. Conclusions
We show that scaled depth to the SMT ( Ls ) can be used to estimate
AGHS for deep-source gas hydrate systems. Simulation results demonstrate that
the net fluid flux controls Ls and the average gas hydrate flux ( Pe1 S h ) through
the GHSZ, thereby allowing us to correlate L and Pe S . Results also show
s 1 h
that conditions that create shallow Ls and low Peclet number ( Pe1 ) lead to higher
AGHS. Application of this method to sites along Cascadia Margin reveals a good
match with saturations estimated from chloride/resistivity log data and accurately
predicts the lateral variability in AGHS.
References
Berner, R. A. (1980), Early Diagenesis : A Theoretical Approach, Princeton Univ.
Press, Princeton, N.J.
Bhatnagar, G., W. G. Chapman, G. R. Dickens, B. Dugan, and G. J. Hirasaki
(2007), Generalization of gas hydrate distribution and saturation in marine
sediments by scaling of thermodynamic and transport processes, Am. J. Sci.,
307, 861-900.
Borowski, W. S., C. K. Paull, and W. Ussler III (1996), Marine pore-water sulfate
profiles indicate in situ methane flux from underlying gas hydrate, Geology,
24(7), 655-658.
Borowski, W. S., C. K. Paull, and W. Ussler III (1999), Global and local variations
of interstitial sulfate gradients in deep-water, continental margin sediments:
Sensitivity to underlying methane and gas hydrates, Mar. Geol., 159, 131-154.
7
Claypool, G. E., and K. A. Kvenvolden (1983), Methane and other hydrocarbon
gases in marine sediment, Annu. Rev. Earth Planet. Sci., 11, 299-327.
Davie, M. K., and B. A. Buffett (2003), Sources of methane for marine gas
hydrate: inferences from a comparison of observations and numerical models,
Earth Planet. Sci. Lett., 206, 51-63.
Dickens, G. R. (2003), Rethinking the global carbon cycle with a large, dynamic
and microbially mediated gas hydrate capacitor, Earth Planet. Sci. Lett., 213,
169-183.
Hyndman, R. D., and E. E. Davis (1992), A mechanism for the formation of
methane hydrate and seafoor bottom-simulating reflectors by vertical fluid
expulsion, J. Geophys. Res. B, 97, 7025-7041.
Hyndman, R. D., T. Yuan, and K. Moran (1999), The concentration of deep sea
gas hydrates from downhole electrical resistivity logs and laboratory data,
Earth Planet. Sci. Lett., 172, 167-177.
Iversen, N., and B. B. Jørgensen (1993), Diffusion coefficients of sulfate and
methane in marine sediments: Influence of porosity, Geochim. Cosmochim.
Acta, 57, 571-578.
Kvenvolden, K. A. (1993), Gas hydrates: Geological perspective and global
change, Rev. Geophys., 31, 173-187.
Riedel, M., T. S. Collett, M. J. Malone, and the Expedition 311 Scientists (Eds.)
(2006), Proc. IODP, vol. 311, Integrated Ocean Drilling Program Management
International Inc., Washington, DC.
Snyder, G. T., A. Hiruta, R. Matsumoto, G. R. Dickens, H. Tomaru, R. Takeuchi,
J. Komatsubara, Y. Ishida, and H. Yu (2007), Pore water profiles and
authigenic mineralization in shallow marine sediments above the methane-
charged system on Umitaka Spur, Japan Sea, Deep-Sea Res. II, 54, 1216-
1239.
Ussler, W., and C. K. Paull (2001), Ion exclusion associated with marine gas
hydrate deposits, in Natural Gas Hydrates: Occurrence, Distribution, and
Detection, Geophys. Monogr. Ser., vol. 124, edited by C. K. Paull and W. P.
Dillon, pp. 41{51, AGU, Washington, D. C.
Westbrook, G. K., B. Carson, and R. J. Musgrave et al. (Eds.) (1994), Proc. ODP,
Initial Reports, vol. 146 (Pt. 1), Ocean Drilling Program, College Statiom, TX.
8
Table 1. Site-specific parameters and calculated AGHS for Cascadia Margin
sites compared with estimates from resistivity log data and chloride
anomalies
Seafloor Ls / Lt Pe1 S h Sh
S Sh Sh
Site Pe1 depth (m/m) =
(cm/k.y.) (Figure (res. (Cl¯)
(m) Ls 4)
(calc.)
log)
10/225
889a 25 0.07 1311 0.03 0.4% - <1% b
= 0.044
4.5/230
U1325c 38.3 0.11 2195 0.22 2% 3.7%e 5.3%f
= 0.02
2.5/230
U1326c 38.3g 0.11 1828 0.46 4.2% 6.7%e 5.5%f
= 0.011
1.5/219
U1328c 34.3 0.09 1267 0.67 7.4% 12.6%e -
= 0.007
a
ODP Leg 146 [Westbrook et al., 1994]
b
Taken from numerical model of Davie and Buffett, 2003
c
IODP Expedition 311 [Riedel et al., 2006]
e
Calculated from Archie equation using resistivity log data [Riedel et al., 2006]
f
Calculated using relative freshening of pore water chloride profiles [Riedel et al.,
2006]
g
S was not available, hence assumed equal to rate at nearest site U1325
9
Figure 1. (A) Schematic representation of a gas hydrate system showing pore
water sulfate and methane concentrations, which go to zero at some shallow
depth because of anaerobic oxidation of methane (AOM). Also shown are
methane solubility in water, the two fluid fluxes ( U f , sed and U f ,ext ), and depth to
the base of the gas hydrate stability zone ( Lt ). (B) Close-up of the sulfate-
methane transition (SMT) showing overlap of sulfate and methane profiles, and
its depth below the seafloor ( Ls ).
10
Figure 2. Effect of Damkohler number ( Da AOM ) and Peclet numbers ( Pe1 , Pe2 ) on
steady-state profiles. Pe1 + Pe2 = −10 for all cases. Note different y-axis scale for
each plot. (a) Sulfate and methane profiles for Da AOM = 108 (solid curves) and
Da AOM = 106 (dashed curves). Hatched regions compare the thickness of the
SMT for the two cases. (b) Simulations for different sets of Pe1 and Pe2 ,
with Pe1 + Pe2 = −10 . Overlap of methane and sulfate profiles shows that Pe1 + Pe2
controls the concentrations. (c) The product Pe1 S h depends on Pe1 + Pe2 , but
hydrate saturation profiles are a function of Pe1 .
11
Figure 3. Effect of net fluid flux ( Pe1 + Pe2 ) on steady-state concentrations. Pe1
equals 0.1 for all simulations. (a) High magnitude of Pe1 + Pe2 defines higher net
methane fluxes, resulting in shallower SMT zones. (b) Gas hydrate saturation at
steady state increases as magnitude of Pe1 + Pe2 increases.
12
Figure 4. Relationship between average gas hydrate flux Pe1 S h and scaled
SMT depth ( Ls = Ls / Lt ) for several seafloor depths. Points corresponding to four
Cascadia Margin sites are plotted to show how AGHS is estimated from L usings
this plot (Table 1).
13
National Energy Technology Laboratory
Customer Service:
1-800-553-7681